diff --git a/.beads/issues.jsonl b/.beads/issues.jsonl
index de274e2..66c9a3c 100644
--- a/.beads/issues.jsonl
+++ b/.beads/issues.jsonl
@@ -2,64 +2,86 @@
 {"id":"bd-10i","title":"Epic: CP2 Gate D - Resumability Proof","description":"## Background\nGate D validates resumability and crash recovery. Proves that cursor and watermark mechanics prevent massive refetch after interruption. This is critical for large projects where a full refetch would take hours.\n\n## Acceptance Criteria (Pass/Fail)\n- [ ] Kill mid-run, rerun -> bounded redo (not full refetch from beginning)\n- [ ] Cursor saved at page boundary (not item boundary)\n- [ ] No redundant discussion refetch after crash recovery\n- [ ] No watermark advancement on partial pagination failure\n- [ ] Single-flight lock prevents concurrent ingest runs\n- [ ] `--full` flag resets MR cursor to NULL\n- [ ] `--full` flag resets ALL `discussions_synced_for_updated_at` to NULL\n- [ ] `--force` bypasses single-flight lock\n\n## Validation Script\n```bash\n#!/bin/bash\nset -e\n\nDB_PATH=\"${XDG_DATA_HOME:-$HOME/.local/share}/gitlab-inbox/db.sqlite3\"\n\necho \"=== Gate D: Resumability Proof ===\"\n\n# 1. Test single-flight lock\necho \"Step 1: Test single-flight lock...\"\ngi ingest --type=merge_requests &\nFIRST_PID=$!\nsleep 1\n\n# Try second ingest - should fail with lock error\nif gi ingest --type=merge_requests 2>&1 | grep -q \"lock\\|already running\"; then\n  echo \"  PASS: Second ingest blocked by lock\"\nelse\n  echo \"  FAIL: Lock not working\"\nfi\nwait $FIRST_PID 2>/dev/null || true\n\n# 2. Test --force bypasses lock\necho \"Step 2: Test --force flag...\"\ngi ingest --type=merge_requests &\nFIRST_PID=$!\nsleep 1\nif gi ingest --type=merge_requests --force 2>&1; then\n  echo \"  PASS: --force bypassed lock\"\nelse\n  echo \"  Note: --force test inconclusive\"\nfi\nwait $FIRST_PID 2>/dev/null || true\n\n# 3. Check cursor state\necho \"Step 3: Check cursor state...\"\nsqlite3 \"$DB_PATH\" \"\n  SELECT resource_type, updated_at, gitlab_id\n  FROM sync_cursors \n  WHERE resource_type = 'merge_requests';\n\"\n\n# 4. Test crash recovery\necho \"Step 4: Test crash recovery...\"\n\n# Record current cursor\nCURSOR_BEFORE=$(sqlite3 \"$DB_PATH\" \"\n  SELECT updated_at FROM sync_cursors WHERE resource_type = 'merge_requests';\n\")\necho \"  Cursor before: $CURSOR_BEFORE\"\n\n# Force full sync and kill\necho \"  Starting full sync then killing...\"\ngi ingest --type=merge_requests --full &\nPID=$!\nsleep 5 && kill -9 $PID 2>/dev/null || true\nwait $PID 2>/dev/null || true\n\n# Check cursor was saved (should be non-null if any page completed)\nCURSOR_AFTER=$(sqlite3 \"$DB_PATH\" \"\n  SELECT updated_at FROM sync_cursors WHERE resource_type = 'merge_requests';\n\")\necho \"  Cursor after kill: $CURSOR_AFTER\"\n\n# Re-run and verify bounded redo\necho \"  Re-running (should resume from cursor)...\"\ntime gi ingest --type=merge_requests\n# Should be faster than first full sync\n\n# 5. Test --full reset\necho \"Step 5: Test --full resets watermarks...\"\n\n# Check watermarks before\nWATERMARKS_BEFORE=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM merge_requests \n  WHERE discussions_synced_for_updated_at IS NOT NULL;\n\")\necho \"  Watermarks set before --full: $WATERMARKS_BEFORE\"\n\n# Record cursor before\nCURSOR_BEFORE_FULL=$(sqlite3 \"$DB_PATH\" \"\n  SELECT updated_at, gitlab_id FROM sync_cursors WHERE resource_type = 'merge_requests';\n\")\necho \"  Cursor before --full: $CURSOR_BEFORE_FULL\"\n\n# Run --full\ngi ingest --type=merge_requests --full\n\n# Check cursor was reset then rebuilt\nCURSOR_AFTER_FULL=$(sqlite3 \"$DB_PATH\" \"\n  SELECT updated_at, gitlab_id FROM sync_cursors WHERE resource_type = 'merge_requests';\n\")\necho \"  Cursor after --full: $CURSOR_AFTER_FULL\"\n\n# Watermarks should be set again (sync completed)\nWATERMARKS_AFTER=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM merge_requests \n  WHERE discussions_synced_for_updated_at IS NOT NULL;\n\")\necho \"  Watermarks set after --full: $WATERMARKS_AFTER\"\n\necho \"\"\necho \"=== Gate D: PASSED ===\"\n```\n\n## Watermark Safety Test (Simulated Network Failure)\n```bash\n# This tests that watermark doesn't advance on partial failure\n# Requires ability to simulate network issues\n\n# 1. Get an MR that needs discussion sync\nMR_ID=$(sqlite3 \"$DB_PATH\" \"\n  SELECT id FROM merge_requests \n  WHERE discussions_synced_for_updated_at IS NULL \n     OR updated_at > discussions_synced_for_updated_at\n  LIMIT 1;\n\")\n\n# 2. Note current watermark\nWATERMARK_BEFORE=$(sqlite3 \"$DB_PATH\" \"\n  SELECT discussions_synced_for_updated_at FROM merge_requests WHERE id = $MR_ID;\n\")\necho \"Watermark before: $WATERMARK_BEFORE\"\n\n# 3. Simulate network failure (requires network manipulation)\n# Option A: Block GitLab API temporarily\n# Option B: Run in a container with network limits\n# Option C: Use the automated test instead:\ncargo test does_not_advance_discussion_watermark_on_partial_failure\n\n# 4. Verify watermark unchanged after failure\nWATERMARK_AFTER=$(sqlite3 \"$DB_PATH\" \"\n  SELECT discussions_synced_for_updated_at FROM merge_requests WHERE id = $MR_ID;\n\")\necho \"Watermark after failure: $WATERMARK_AFTER\"\n[ \"$WATERMARK_BEFORE\" = \"$WATERMARK_AFTER\" ] && echo \"PASS: Watermark preserved\"\n```\n\n## Test Commands (Quick Verification)\n```bash\n# Check cursor state:\nsqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"\n  SELECT * FROM sync_cursors WHERE resource_type = 'merge_requests';\n\"\n\n# Check watermark distribution:\nsqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"\n  SELECT \n    SUM(CASE WHEN discussions_synced_for_updated_at IS NULL THEN 1 ELSE 0 END) as needs_sync,\n    SUM(CASE WHEN discussions_synced_for_updated_at IS NOT NULL THEN 1 ELSE 0 END) as synced\n  FROM merge_requests;\n\"\n\n# Test --full resets (check before/after):\nsqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"SELECT COUNT(*) FROM merge_requests WHERE discussions_synced_for_updated_at IS NOT NULL;\"\ngi ingest --type=merge_requests --full\n# During full sync, watermarks should be NULL, then repopulated\n```\n\n## Critical Automated Tests\nThese tests MUST pass for Gate D:\n```bash\ncargo test does_not_advance_discussion_watermark_on_partial_failure\ncargo test full_sync_resets_discussion_watermarks\ncargo test cursor_saved_at_page_boundary\n```\n\n## Dependencies\nThis gate requires:\n- bd-mk3 (ingest command with --full and --force support)\n- bd-ser (MR ingestion with cursor mechanics)\n- bd-20h (MR discussion ingestion with watermark safety)\n- Gates A, B, C must pass first\n\n## Edge Cases\n- Very fast sync: May complete before kill signal reaches; retest with larger project\n- Lock file stale: If previous run crashed, lock file may exist; --force handles this\n- Clock skew: Cursor timestamps should use server time, not local time","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-26T22:06:02.124186Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:48:21.060596Z","closed_at":"2026-01-27T00:48:21.060555Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-10i","depends_on_id":"bd-mk3","type":"blocks","created_at":"2026-01-26T22:08:55.875790Z","created_by":"tayloreernisse"}]}
 {"id":"bd-13b","title":"[CP0] CLI entry point with Commander.js","description":"## Background\n\nCommander.js provides the CLI framework. The main entry point sets up the program with all subcommands. Uses ESM with proper shebang for npx/global installation.\n\nReference: docs/prd/checkpoint-0.md section \"CLI Commands\"\n\n## Approach\n\n**src/cli/index.ts:**\n```typescript\n#!/usr/bin/env node\n\nimport { Command } from 'commander';\nimport { version } from '../../package.json' with { type: 'json' };\nimport { initCommand } from './commands/init';\nimport { authTestCommand } from './commands/auth-test';\nimport { doctorCommand } from './commands/doctor';\nimport { versionCommand } from './commands/version';\nimport { backupCommand } from './commands/backup';\nimport { resetCommand } from './commands/reset';\nimport { syncStatusCommand } from './commands/sync-status';\n\nconst program = new Command();\n\nprogram\n  .name('gi')\n  .description('GitLab Inbox - Unified notification management')\n  .version(version);\n\n// Global --config flag available to all commands\nprogram.option('-c, --config <path>', 'Path to config file');\n\n// Register subcommands\nprogram.addCommand(initCommand);\nprogram.addCommand(authTestCommand);\nprogram.addCommand(doctorCommand);\nprogram.addCommand(versionCommand);\nprogram.addCommand(backupCommand);\nprogram.addCommand(resetCommand);\nprogram.addCommand(syncStatusCommand);\n\nprogram.parse();\n```\n\nEach command file exports a Command instance:\n```typescript\n// src/cli/commands/version.ts\nimport { Command } from 'commander';\n\nexport const versionCommand = new Command('version')\n  .description('Show version information')\n  .action(() => {\n    console.log(`gi version ${version}`);\n  });\n```\n\n## Acceptance Criteria\n\n- [ ] `gi --help` shows all commands and global options\n- [ ] `gi --version` shows version from package.json\n- [ ] `gi <command> --help` shows command-specific help\n- [ ] `gi --config ./path` passes config path to commands\n- [ ] Unknown command shows error and suggests --help\n- [ ] Exit code 0 on success, non-zero on error\n- [ ] Shebang line works for npx execution\n\n## Files\n\nCREATE:\n- src/cli/index.ts (main entry point)\n- src/cli/commands/version.ts (simple command as template)\n\nMODIFY (later beads):\n- package.json (add \"bin\" field pointing to dist/cli/index.js)\n\n## TDD Loop\n\nN/A for CLI entry point - verify with manual testing:\n\n```bash\nnpm run build\nnode dist/cli/index.js --help\nnode dist/cli/index.js version\nnode dist/cli/index.js unknown-command  # should error\n```\n\n## Edge Cases\n\n- package.json import requires Node 20+ with { type: 'json' } assertion\n- Alternative: read version from package.json with readFileSync\n- Command registration order affects help display - alphabetical preferred\n- Global options must be defined before subcommands","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:50.499023Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:10:49.224627Z","closed_at":"2026-01-25T03:10:49.224499Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-13b","depends_on_id":"bd-gg1","type":"blocks","created_at":"2026-01-24T16:13:09.370408Z","created_by":"tayloreernisse"}]}
 {"id":"bd-140","title":"[CP1] Database migration 002_issues.sql","description":"Create migration file with tables for issues, labels, issue_labels, discussions, and notes.\n\nTables to create:\n- issues: gitlab_id, project_id, iid, title, description, state, author_username, timestamps, web_url, raw_payload_id\n- labels: gitlab_id, project_id, name, color, description (unique on project_id+name)\n- issue_labels: junction table\n- discussions: gitlab_discussion_id, project_id, issue_id, noteable_type, individual_note, timestamps, resolvable/resolved\n- notes: gitlab_id, discussion_id, project_id, type, is_system, author_username, body, timestamps, position, resolution fields, DiffNote position fields\n\nInclude appropriate indexes:\n- idx_issues_project_updated, idx_issues_author, uq_issues_project_iid\n- uq_labels_project_name, idx_labels_name\n- idx_issue_labels_label\n- uq_discussions_project_discussion_id, idx_discussions_issue/mr/last_note\n- idx_notes_discussion/author/system\n\nFiles: migrations/002_issues.sql\nDone when: Migration applies cleanly on top of 001_initial.sql","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:18:53.954039Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.154936Z","deleted_at":"2026-01-25T15:21:35.154934Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-14q","title":"Epic: Gate 4 - File Decision History (lore file-history)","description":"Implement 'lore file-history <path>' command showing MRs that touched a file, with rename chain resolution, discussion context, and linked issues. Includes migration 012 (renumbered from spec's 011) for mr_file_changes and commit SHA fields.\n\nChildren: bd-1oo (migration 012), bd-jec (config), bd-2yo (MR diffs API), bd-1yx (rename chains), bd-z94 (file-history command)","status":"open","priority":1,"issue_type":"feature","created_at":"2026-02-02T21:31:01.094024Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:51.656482Z","compaction_level":0,"original_size":0,"labels":["epic","gate-4","phase-b"],"dependencies":[{"issue_id":"bd-14q","depends_on_id":"bd-1se","type":"blocks","created_at":"2026-02-02T21:34:16.913465Z","created_by":"tayloreernisse"},{"issue_id":"bd-14q","depends_on_id":"bd-2zl","type":"blocks","created_at":"2026-02-02T21:34:16.870058Z","created_by":"tayloreernisse"}]}
 {"id":"bd-157","title":"[CP1] Issue transformer with label extraction","description":"Transform GitLab issue payloads to normalized database schema.\n\n## Module\nsrc/gitlab/transformers/issue.rs\n\n## Structs\n\n### NormalizedIssue\n- gitlab_id: i64\n- project_id: i64 (local DB project ID)\n- iid: i64\n- title: String\n- description: Option<String>\n- state: String\n- author_username: String\n- created_at, updated_at, last_seen_at: i64 (ms epoch)\n- web_url: String\n\n### NormalizedLabel (CP1: name-only)\n- project_id: i64\n- name: String\n\n## Functions\n\n### transform_issue(gitlab_issue: &GitLabIssue, local_project_id: i64) -> NormalizedIssue\n- Convert ISO timestamps to ms epoch using iso_to_ms()\n- Set last_seen_at to now_ms()\n- Clone string fields\n\n### extract_labels(gitlab_issue: &GitLabIssue, local_project_id: i64) -> Vec<NormalizedLabel>\n- Map labels vec to NormalizedLabel structs\n\nFiles: \n- src/gitlab/transformers/mod.rs\n- src/gitlab/transformers/issue.rs\nTests: tests/issue_transformer_tests.rs\nDone when: Unit tests pass for payload transformation and label extraction","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:42:47.719562Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.736142Z","deleted_at":"2026-01-25T17:02:01.736129Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-17v","title":"[CP1] gi sync-status enhancement","description":"## Background\n\nThe `gi sync-status` command shows synchronization state: last successful sync time, cursor positions per project/resource, and overall health. This helps users understand when data was last refreshed and diagnose sync issues.\n\n## Approach\n\n### Module: src/cli/commands/sync_status.rs (enhance existing or create)\n\n### Handler Function\n\n```rust\npub async fn handle_sync_status(conn: &Connection) -> Result<()>\n```\n\n### Data to Display\n\n1. **Last sync run**: From `sync_runs` table\n   - Started at, completed at, status\n   - Issues fetched, discussions fetched\n\n2. **Cursor positions**: From `sync_cursors` table\n   - Per (project, resource_type) pair\n   - Show updated_at_cursor as human-readable date\n   - Show tie_breaker_id (GitLab ID of last processed item)\n\n3. **Overall counts**: Quick summary\n   - Total issues, discussions, notes in DB\n\n### Output Format\n\n```\nLast Sync\n─────────\nStatus:    completed\nStarted:   2024-01-25 10:30:00\nCompleted: 2024-01-25 10:35:00\nDuration:  5m 23s\n\nCursor Positions\n────────────────\ngroup/project-one (issues):\n  Last updated_at: 2024-01-25 10:30:00\n  Last GitLab ID:  12345\n\nData Summary\n────────────\nIssues:      1,234\nDiscussions: 5,678\nNotes:       12,345 (excluding 2,000 system)\n```\n\n### Queries\n\n```sql\n-- Last sync run\nSELECT * FROM sync_runs ORDER BY started_at DESC LIMIT 1\n\n-- Cursor positions\nSELECT p.path, sc.resource_type, sc.updated_at_cursor, sc.tie_breaker_id\nFROM sync_cursors sc\nJOIN projects p ON sc.project_id = p.id\n\n-- Data summary\nSELECT COUNT(*) FROM issues\nSELECT COUNT(*) FROM discussions\nSELECT COUNT(*), SUM(is_system) FROM notes\n```\n\n## Acceptance Criteria\n\n- [ ] Shows last sync run with status and timing\n- [ ] Shows cursor position per project/resource\n- [ ] Shows total counts for issues, discussions, notes\n- [ ] Handles case where no sync has run yet\n- [ ] Formats timestamps as human-readable local time\n\n## Files\n\n- src/cli/commands/sync_status.rs (create or enhance)\n- src/cli/mod.rs (add SyncStatus variant if new)\n\n## TDD Loop\n\nRED:\n```rust\n#[tokio::test] async fn sync_status_shows_last_run()\n#[tokio::test] async fn sync_status_shows_cursor_positions()\n#[tokio::test] async fn sync_status_handles_no_sync_yet()\n```\n\nGREEN: Implement handler with queries and formatting\n\nVERIFY: `cargo test sync_status`\n\n## Edge Cases\n\n- No sync has ever run - show \"No sync runs recorded\"\n- Sync in progress - show \"Status: running\" with started_at\n- Cursor at epoch 0 - means fresh start, show \"Not started\"\n- Multiple projects - show cursor for each","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-25T17:02:38.409353Z","created_by":"tayloreernisse","updated_at":"2026-01-25T23:03:21.851557Z","closed_at":"2026-01-25T23:03:21.851496Z","close_reason":"Implemented gi sync-status showing last run, cursor positions, and data summary","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-17v","depends_on_id":"bd-208","type":"blocks","created_at":"2026-01-25T17:04:05.749433Z","created_by":"tayloreernisse"}]}
 {"id":"bd-18t","title":"Implement discussion truncation logic","description":"## Background\nDiscussion threads can contain dozens of notes spanning thousands of characters. The truncation module ensures discussion documents stay within a 32k character limit (suitable for embedding chunking) by dropping middle notes while preserving first and last notes for context. A separate hard safety cap of 2MB applies to ALL document types for pathological content (pasted logs, base64 blobs). Issue/MR documents are NOT truncated by the discussion logic — only the hard cap applies.\n\n## Approach\nCreate `src/documents/truncation.rs` per PRD Section 2.3:\n\n```rust\npub const MAX_DISCUSSION_CHARS: usize = 32_000;\npub const MAX_DOCUMENT_CHARS_HARD: usize = 2_000_000;\n\npub struct NoteContent {\n    pub author: String,\n    pub date: String,\n    pub body: String,\n}\n\npub struct TruncationResult {\n    pub content: String,\n    pub is_truncated: bool,\n    pub reason: Option<TruncationReason>,\n}\n\npub enum TruncationReason {\n    TokenLimitMiddleDrop,\n    SingleNoteOversized,\n    FirstLastOversized,\n    HardCapOversized,\n}\n```\n\n**Core functions:**\n- `truncate_discussion(notes: &[NoteContent], max_chars: usize) -> TruncationResult`\n- `truncate_utf8(s: &str, max_bytes: usize) -> &str` (shared with fts.rs)\n- `truncate_hard_cap(content: &str) -> TruncationResult` (for any doc type)\n\n**Algorithm for truncate_discussion:**\n1. Format all notes as `@author (date):\\nbody\\n\\n`\n2. If total <= max_chars: return as-is\n3. If single note: truncate at UTF-8 boundary, append `[truncated]`, reason = SingleNoteOversized\n4. Binary search: find max N where first N notes + last 1 note + marker fit within max_chars\n5. If first + last > max_chars: keep only first (truncated), reason = FirstLastOversized\n6. Otherwise: first N + marker + last M, reason = TokenLimitMiddleDrop\n\n**Marker format:** `\\n\\n[... N notes omitted for length ...]\\n\\n`\n\n## Acceptance Criteria\n- [ ] Discussion with total < 32k chars returns untruncated\n- [ ] Discussion > 32k chars: middle notes dropped, first + last preserved\n- [ ] Truncation marker shows correct count of omitted notes\n- [ ] Single note > 32k chars: truncated at UTF-8-safe boundary with `[truncated]` appended\n- [ ] First + last note > 32k: only first note kept (truncated if needed)\n- [ ] Hard cap (2MB) truncates any document type at UTF-8-safe boundary\n- [ ] `truncate_utf8` never panics on multi-byte codepoints (emoji, CJK, accented chars)\n- [ ] `TruncationReason::as_str()` returns DB-compatible strings matching CHECK constraint\n\n## Files\n- `src/documents/truncation.rs` — new file\n- `src/documents/mod.rs` — add `pub use truncation::{truncate_discussion, truncate_hard_cap, TruncationResult, NoteContent};`\n\n## TDD Loop\nRED: Tests in `#[cfg(test)] mod tests`:\n- `test_no_truncation_under_limit` — 3 short notes, all fit\n- `test_middle_notes_dropped` — 10 notes totaling > 32k, first+last preserved\n- `test_single_note_oversized` — one note of 50k chars, truncated safely\n- `test_first_last_oversized` — first=20k, last=20k, only first kept\n- `test_one_note_total` — single note under limit: no truncation\n- `test_utf8_boundary_safety` — content with emoji/CJK at truncation point\n- `test_hard_cap` — 3MB content truncated to 2MB\n- `test_marker_count_correct` — marker says \"[... 5 notes omitted ...]\" when 5 dropped\nGREEN: Implement truncation logic\nVERIFY: `cargo test truncation`\n\n## Edge Cases\n- Empty notes list: return empty content, not truncated\n- All notes are empty strings: total = 0, no truncation\n- Note body contains only multi-byte characters: truncate_utf8 walks backward to find safe boundary\n- Note body with trailing newlines: formatted output should not have excessive blank lines","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:25:45.597167Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:21:32.256569Z","closed_at":"2026-01-30T17:21:32.256507Z","close_reason":"Completed: truncate_discussion, truncate_hard_cap, truncate_utf8, TruncationReason with as_str(), 12 tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-18t","depends_on_id":"bd-36p","type":"blocks","created_at":"2026-01-30T15:29:15.947679Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1cb","title":"[CP0] gi doctor command - health checks","description":"## Background\n\ndoctor is the primary diagnostic command. It checks all system components and reports their status. Supports JSON output for scripting and CI integration. Must degrade gracefully - warn about optional components (Ollama) without failing.\n\nReference: docs/prd/checkpoint-0.md section \"gi doctor\"\n\n## Approach\n\n**src/cli/commands/doctor.ts:**\n\nPerforms 5 checks:\n1. **Config**: Load and validate config file\n2. **Database**: Open DB, verify pragmas, check schema version\n3. **GitLab**: Auth with token, verify connectivity\n4. **Projects**: Count configured vs resolved in DB\n5. **Ollama**: Ping embedding endpoint (optional - warn if unavailable)\n\n**DoctorResult interface:**\n```typescript\ninterface DoctorResult {\n  success: boolean;  // All required checks passed\n  checks: {\n    config: { status: 'ok' | 'error'; path?: string; error?: string };\n    database: { status: 'ok' | 'error'; path?: string; schemaVersion?: number; error?: string };\n    gitlab: { status: 'ok' | 'error'; url?: string; username?: string; error?: string };\n    projects: { status: 'ok' | 'error'; configured?: number; resolved?: number; error?: string };\n    ollama: { status: 'ok' | 'warning' | 'error'; url?: string; model?: string; error?: string };\n  };\n}\n```\n\n**Human-readable output (default):**\n```\ngi doctor\n\n  Config     ✓  Loaded from ~/.config/gi/config.json\n  Database   ✓  ~/.local/share/gi/data.db (schema v1)\n  GitLab     ✓  https://gitlab.example.com (authenticated as @johndoe)\n  Projects   ✓  2 configured, 2 resolved\n  Ollama     ⚠  Not running (semantic search unavailable)\n\nStatus: Ready (lexical search available, semantic search requires Ollama)\n```\n\n**JSON output (--json flag):**\nOutputs DoctorResult as JSON to stdout\n\n## Acceptance Criteria\n\n- [ ] Config check: shows path and validation status\n- [ ] Database check: shows path, schema version, pragma verification\n- [ ] GitLab check: shows URL and authenticated username\n- [ ] Projects check: shows configured count and resolved count\n- [ ] Ollama check: warns if not running, doesn't fail overall\n- [ ] success=true only if config, database, gitlab, projects all ok\n- [ ] --json outputs valid JSON matching DoctorResult interface\n- [ ] Exit 0 if success=true, exit 1 if any required check fails\n- [ ] Colors and symbols in human output (✓, ⚠, ✗)\n\n## Files\n\nCREATE:\n- src/cli/commands/doctor.ts\n- src/types/doctor.ts (DoctorResult interface)\n\n## TDD Loop\n\nN/A - diagnostic command, verify with manual testing:\n\n```bash\n# All good\ngi doctor\n\n# JSON output\ngi doctor --json | jq .\n\n# With missing Ollama\n# (just don't run Ollama - should show warning)\n\n# With bad config\nmv ~/.config/gi/config.json ~/.config/gi/config.json.bak\ngi doctor  # should show config error\n```\n\n## Edge Cases\n\n- Ollama timeout should be short (2s) - don't block on slow network\n- Ollama 404 (wrong model) vs connection refused (not running)\n- Database file exists but wrong schema version\n- Projects in config but not in database (init not run)\n- Token valid for user but project access revoked","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:51.435540Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:30:24.921206Z","closed_at":"2026-01-25T03:30:24.921041Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1cb","depends_on_id":"bd-13b","type":"blocks","created_at":"2026-01-24T16:13:10.427307Z","created_by":"tayloreernisse"},{"issue_id":"bd-1cb","depends_on_id":"bd-1l1","type":"blocks","created_at":"2026-01-24T16:13:10.478469Z","created_by":"tayloreernisse"},{"issue_id":"bd-1cb","depends_on_id":"bd-3ng","type":"blocks","created_at":"2026-01-24T16:13:10.461940Z","created_by":"tayloreernisse"},{"issue_id":"bd-1cb","depends_on_id":"bd-epj","type":"blocks","created_at":"2026-01-24T16:13:10.443612Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1d5","title":"[CP1] GitLab client pagination methods","description":"Add async generator methods for paginated GitLab API calls.\n\nMethods to add to src/gitlab/client.ts:\n- paginateIssues(gitlabProjectId, updatedAfter?) → AsyncGenerator<GitLabIssue>\n- paginateIssueDiscussions(gitlabProjectId, issueIid) → AsyncGenerator<GitLabDiscussion>\n- requestWithHeaders<T>(path) → { data: T, headers: Headers }\n\nImplementation:\n- Use scope=all, state=all for issues\n- Order by updated_at ASC\n- Follow X-Next-Page header until empty/absent\n- Apply cursor rewind (subtract cursorRewindSeconds) for tuple semantics\n- Fall back to empty-page detection if headers missing\n\nFiles: src/gitlab/client.ts\nTests: tests/unit/pagination.test.ts\nDone when: Pagination handles multiple pages and respects cursors","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:19:43.069869Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.156881Z","deleted_at":"2026-01-25T15:21:35.156877Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-1ep","title":"Wire resource event fetching into sync pipeline","description":"## Background\nAfter issue/MR primary ingestion and discussion fetch, changed entities need resource_events jobs enqueued and drained. This is the integration point that connects the queue (bd-tir), API client (bd-sqw), DB upserts (bd-1uc), and config flag (bd-2e8).\n\n## Approach\nModify the sync pipeline to add two new phases after discussion sync:\n\n**Phase 1 — Enqueue during ingestion:**\nIn src/ingestion/orchestrator.rs, after each entity upsert (issue or MR), call:\n```rust\nif config.sync.fetch_resource_events {\n    enqueue_job(conn, project_id, \"issue\", iid, local_id, \"resource_events\", None)?;\n}\n// For MRs, also enqueue mr_closes_issues (always) and mr_diffs (when fetchMrFileChanges)\n```\n\nThe \"changed entity\" detection uses the existing dirty tracker: if an entity was inserted or updated during this sync run, it gets enqueued. On --full sync, all entities are enqueued.\n\n**Phase 2 — Drain dependent queue:**\nAdd a new drain step in src/cli/commands/sync.rs (or new src/core/drain.rs), called after discussion sync:\n```rust\npub async fn drain_dependent_queue(\n    conn: &Connection,\n    client: &GitLabClient,\n    config: &Config,\n    progress: Option<ProgressCallback>,\n) -> Result<DrainResult>\n```\n\nFlow:\n1. reclaim_stale_locks(conn, config.sync.stale_lock_minutes)\n2. Loop: claim_jobs(conn, \"resource_events\", batch_size=10)\n3. For each job:\n   a. Fetch 3 event types via client (fetch_issue_state_events etc.)\n   b. Store via upsert functions (upsert_state_events etc.)\n   c. complete_job(conn, job.id) on success\n   d. fail_job(conn, job.id, error_msg) on failure\n4. Report progress: \"Fetching resource events... [N/M]\"\n5. Repeat until no more claimable jobs\n\n**Progress reporting:**\nAdd new ProgressEvent variants:\n```rust\nResourceEventsFetchStart { total: usize },\nResourceEventsFetchProgress { completed: usize, total: usize },\nResourceEventsFetchComplete { fetched: usize, failed: usize },\n```\n\n## Acceptance Criteria\n- [ ] Full sync enqueues resource_events jobs for all issues and MRs\n- [ ] Incremental sync only enqueues for entities changed since last sync\n- [ ] --no-events prevents enqueueing resource_events jobs\n- [ ] Drain step fetches all 3 event types per entity\n- [ ] Successful fetches stored and job completed\n- [ ] Failed fetches recorded with error, job retried on next sync\n- [ ] Stale locks reclaimed at drain start\n- [ ] Progress displayed: \"Fetching resource events... [N/M]\"\n- [ ] Robot mode progress suppressed (quiet mode)\n\n## Files\n- src/ingestion/orchestrator.rs (add enqueue calls during upsert)\n- src/cli/commands/sync.rs (add drain step after discussions)\n- src/core/drain.rs (new, optional — or inline in sync.rs)\n\n## TDD Loop\nRED: tests/sync_pipeline_tests.rs (or extend existing):\n- `test_sync_enqueues_resource_events_for_changed_entities` - mock sync, verify jobs enqueued\n- `test_sync_no_events_flag_skips_enqueue` - verify no jobs when flag false\n- `test_drain_completes_jobs_on_success` - mock API responses, verify jobs deleted\n- `test_drain_fails_jobs_on_error` - mock API failure, verify job attempts incremented\n\nNote: Full pipeline integration tests may need mock HTTP server. Start with unit tests on enqueue/drain logic using the real DB with mock API responses.\n\nGREEN: Implement enqueue hooks + drain step\n\nVERIFY: `cargo test sync -- --nocapture && cargo build`\n\n## Edge Cases\n- Entity deleted between enqueue and drain: API returns 404, fail_job with \"entity not found\" (retry won't help but backoff caps it)\n- Rate limiting during drain: GitLabRateLimited error should fail_job with retry (transient)\n- Network error during drain: GitLabNetworkError should fail_job with retry\n- Multiple sync runs competing: locked_at prevents double-processing; stale lock reclaim handles crashes\n- Drain should have a max iterations guard to prevent infinite loop if jobs keep failing and being retried within the same run","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:57.334527Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:43:04.365143Z","compaction_level":0,"original_size":0,"labels":["gate-1","phase-b","pipeline"],"dependencies":[{"issue_id":"bd-1ep","depends_on_id":"bd-1uc","type":"blocks","created_at":"2026-02-02T21:32:06.225837Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ep","depends_on_id":"bd-2e8","type":"blocks","created_at":"2026-02-02T21:32:06.142442Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ep","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:57.335847Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ep","depends_on_id":"bd-sqw","type":"blocks","created_at":"2026-02-02T21:32:06.183287Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ep","depends_on_id":"bd-tir","type":"blocks","created_at":"2026-02-02T21:32:06.267800Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1fn","title":"[CP1] Integration tests for discussion watermark","description":"Integration tests verifying discussion sync watermark behavior.\n\n## Tests (tests/discussion_watermark_tests.rs)\n\n- skips_discussion_fetch_when_updated_at_unchanged\n- fetches_discussions_when_updated_at_advanced\n- updates_watermark_after_successful_discussion_sync\n- does_not_update_watermark_on_discussion_sync_failure\n\n## Test Scenario\n1. Ingest issue with updated_at = T1\n2. Verify discussions_synced_for_updated_at = T1\n3. Re-run ingest with same issue (updated_at = T1)\n4. Verify NO discussion API calls made (watermark prevents)\n5. Simulate issue update (updated_at = T2)\n6. Re-run ingest\n7. Verify discussion API calls made for T2\n8. Verify watermark updated to T2\n\n## Why This Matters\nDiscussion API is expensive (1 call per issue). Watermark ensures\nwe only refetch when issue actually changed, even with cursor rewind.\n\nFiles: tests/discussion_watermark_tests.rs\nDone when: Watermark correctly prevents redundant discussion refetch","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:59:11.362495Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:02.086158Z","deleted_at":"2026-01-25T17:02:02.086154Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-1gu","title":"[CP0] gi auth-test command","description":"## Background\n\nauth-test is a quick diagnostic command to verify GitLab connectivity. Used for troubleshooting and CI pipelines. Simpler than doctor because it only checks auth, not full system health.\n\nReference: docs/prd/checkpoint-0.md section \"gi auth-test\"\n\n## Approach\n\n**src/cli/commands/auth-test.ts:**\n```typescript\nimport { Command } from 'commander';\nimport { loadConfig } from '../../core/config';\nimport { GitLabClient } from '../../gitlab/client';\nimport { TokenNotSetError } from '../../core/errors';\n\nexport const authTestCommand = new Command('auth-test')\n  .description('Verify GitLab authentication')\n  .action(async (options, command) => {\n    const globalOpts = command.optsWithGlobals();\n    \n    // 1. Load config\n    const config = loadConfig(globalOpts.config);\n    \n    // 2. Get token from environment\n    const token = process.env[config.gitlab.tokenEnvVar];\n    if (!token) {\n      throw new TokenNotSetError(config.gitlab.tokenEnvVar);\n    }\n    \n    // 3. Create client and test auth\n    const client = new GitLabClient({\n      baseUrl: config.gitlab.baseUrl,\n      token,\n    });\n    \n    // 4. Get current user\n    const user = await client.getCurrentUser();\n    \n    // 5. Output success\n    console.log(`Authenticated as @${user.username} (${user.name})`);\n    console.log(`GitLab: ${config.gitlab.baseUrl}`);\n  });\n```\n\n**Output format:**\n```\nAuthenticated as @johndoe (John Doe)\nGitLab: https://gitlab.example.com\n```\n\n## Acceptance Criteria\n\n- [ ] Loads config from default or --config path\n- [ ] Gets token from configured env var (default GITLAB_TOKEN)\n- [ ] Throws TokenNotSetError if env var not set\n- [ ] Calls GET /api/v4/user to verify auth\n- [ ] Prints username and display name on success\n- [ ] Exit 0 on success\n- [ ] Exit 1 on auth failure (GitLabAuthError)\n- [ ] Exit 1 if config not found (ConfigNotFoundError)\n\n## Files\n\nCREATE:\n- src/cli/commands/auth-test.ts\n\n## TDD Loop\n\nN/A - simple command, verify manually and with integration test in init.test.ts\n\n```bash\n# Manual verification\nexport GITLAB_TOKEN=\"valid-token\"\ngi auth-test\n\n# With invalid token\nexport GITLAB_TOKEN=\"invalid\"\ngi auth-test  # should exit 1\n```\n\n## Edge Cases\n\n- Config exists but token env var not set - clear error message\n- Token exists but wrong scopes - GitLabAuthError (401)\n- Network unreachable - GitLabNetworkError\n- Token with extra whitespace - should trim","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:51.135580Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:28:16.369542Z","closed_at":"2026-01-25T03:28:16.369481Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1gu","depends_on_id":"bd-13b","type":"blocks","created_at":"2026-01-24T16:13:10.058655Z","created_by":"tayloreernisse"},{"issue_id":"bd-1gu","depends_on_id":"bd-1l1","type":"blocks","created_at":"2026-01-24T16:13:10.077581Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1hj","title":"[CP1] Ingestion orchestrator","description":"Coordinate issue + dependent discussion sync with bounded concurrency.\n\n## Module\nsrc/ingestion/orchestrator.rs\n\n## Canonical Pattern (CP1)\n\nWhen gi ingest --type=issues runs:\n\n1. **Ingest issues** - cursor-based with incremental cursor updates per page\n2. **Collect touched issues** - record IssueForDiscussionSync for each issue passing cursor filter\n3. **Filter for discussion sync** - enqueue issues where:\n   issue.updated_at > issues.discussions_synced_for_updated_at\n4. **Execute discussion sync** - with bounded concurrency (dependent_concurrency from config)\n5. **Update watermark** - after each issue's discussions successfully ingested\n\n## Concurrency Notes\n\nRuntime decision: Use single-threaded Tokio runtime (flavor = \"current_thread\")\n- rusqlite::Connection is !Send, conflicts with multi-threaded runtimes\n- Single-threaded avoids Send bounds entirely\n- Use tokio::task::spawn_local + LocalSet for concurrent discussion fetches\n- Keeps code simple; can upgrade to channel-based DB writer in CP2 if needed\n\n## Configuration Used\n- config.sync.dependent_concurrency - limits parallel discussion requests\n- config.sync.cursor_rewind_seconds - safety margin for cursor\n\n## Progress Reporting\n- Show total issues fetched\n- Show issues needing discussion sync\n- Show discussion/note counts per project\n\nFiles: src/ingestion/orchestrator.rs\nTests: Integration tests with mocked GitLab\nDone when: Full issue + discussion ingestion orchestrated correctly","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T16:57:57.325679Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.851047Z","deleted_at":"2026-01-25T17:02:01.851043Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-1ht","title":"Epic: Gate 5 - Code Trace (lore trace)","description":"Implement 'lore trace <file>' command (Tier 1 API-only) that traces file -> MR -> issue -> discussion chains. Builds on file-history (Gate 4) and cross-reference data (Gate 2) to answer 'Why was this code introduced?' Tier 2 (git blame integration) is deferred to Phase C.\n\nChildren: bd-2n4 (trace query), bd-9dd (trace command with human/robot output)","status":"open","priority":1,"issue_type":"feature","created_at":"2026-02-02T21:31:01.141053Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:54.487684Z","compaction_level":0,"original_size":0,"labels":["epic","gate-5","phase-b"],"dependencies":[{"issue_id":"bd-1ht","depends_on_id":"bd-14q","type":"blocks","created_at":"2026-02-02T21:34:38.033428Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ht","depends_on_id":"bd-1se","type":"blocks","created_at":"2026-02-02T21:34:37.987232Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1i2","title":"Integrate mark_dirty_tx into ingestion modules","description":"## Background\nThis bead integrates dirty source tracking into the existing ingestion pipelines. Every entity upserted during ingestion must be marked dirty so the document regenerator knows to update the corresponding search document. The critical constraint: mark_dirty_tx() must be called INSIDE the same transaction that upserts the entity — not after commit.\n\n**Key PRD clarification:** Mark ALL upserted entities dirty (not just changed ones). The regenerator's hash comparison handles \"unchanged\" detection cheaply — this avoids needing change detection in ingestion.\n\n## Approach\nModify 4 existing ingestion files to add mark_dirty_tx() calls inside existing transaction blocks per PRD Section 6.1.\n\n**1. src/ingestion/issues.rs:**\nInside the issue upsert loop, after each successful INSERT/UPDATE:\n```rust\ndirty_tracker::mark_dirty_tx(&tx, SourceType::Issue, issue_row.id)?;\n```\n\n**2. src/ingestion/merge_requests.rs:**\nInside the MR upsert loop:\n```rust\ndirty_tracker::mark_dirty_tx(&tx, SourceType::MergeRequest, mr_row.id)?;\n```\n\n**3. src/ingestion/discussions.rs:**\nInside discussion insert (issue discussions, full-refresh transaction):\n```rust\ndirty_tracker::mark_dirty_tx(&tx, SourceType::Discussion, discussion_row.id)?;\n```\n\n**4. src/ingestion/mr_discussions.rs:**\nInside discussion upsert (write phase):\n```rust\ndirty_tracker::mark_dirty_tx(&tx, SourceType::Discussion, discussion_row.id)?;\n```\n\n**Discussion Sweep Cleanup (PRD Section 6.1 — CRITICAL):**\nWhen the MR discussion sweep deletes stale discussions (`last_seen_at < run_start_time`), **delete the corresponding document rows directly** — do NOT use the dirty queue for cleanup. The `ON DELETE CASCADE` on `document_labels`/`document_paths` and the `documents_embeddings_ad` trigger handle all downstream cleanup.\n\n**PRD-exact CTE pattern:**\n```sql\n-- In src/ingestion/mr_discussions.rs, during sweep phase.\n-- Uses a CTE to capture stale IDs atomically before cascading deletes.\n-- This is more defensive than two separate statements because the CTE\n-- guarantees the ID set is captured before any row is deleted.\nWITH stale AS (\n  SELECT id FROM discussions\n  WHERE merge_request_id = ? AND last_seen_at < ?\n)\n-- Step 1: delete orphaned documents (must happen while source_id still resolves)\nDELETE FROM documents\n  WHERE source_type = 'discussion' AND source_id IN (SELECT id FROM stale);\n-- Step 2: delete the stale discussions themselves\nDELETE FROM discussions\n  WHERE id IN (SELECT id FROM stale);\n```\n\n**NOTE:** If SQLite version doesn't support CTE-based multi-statement, execute as two sequential statements capturing IDs in Rust first:\n```rust\nlet stale_ids: Vec<i64> = conn.prepare(\n    \"SELECT id FROM discussions WHERE merge_request_id = ? AND last_seen_at < ?\"\n)?.query_map(params![mr_id, run_start], |r| r.get(0))?\n  .collect::<Result<Vec<_>, _>>()?;\n\nif !stale_ids.is_empty() {\n    // Delete documents FIRST (while source_id still resolves)\n    conn.execute(\n        \"DELETE FROM documents WHERE source_type = 'discussion' AND source_id IN (...)\",\n        ...\n    )?;\n    // Then delete the discussions\n    conn.execute(\n        \"DELETE FROM discussions WHERE id IN (...)\",\n        ...\n    )?;\n}\n```\n\n**IMPORTANT difference from dirty queue pattern:** The sweep deletes documents DIRECTLY (not via dirty_sources queue). This is because the source entity is being deleted — there's nothing for the regenerator to regenerate from. The cascade handles FTS, labels, paths, and embeddings cleanup.\n\n## Acceptance Criteria\n- [ ] Every upserted issue is marked dirty inside the same transaction\n- [ ] Every upserted MR is marked dirty inside the same transaction\n- [ ] Every upserted discussion (issue + MR) is marked dirty inside the same transaction\n- [ ] ALL upserted entities marked dirty (not just changed ones) — regenerator handles skip\n- [ ] mark_dirty_tx called with &Transaction (not &Connection)\n- [ ] mark_dirty_tx uses upsert with ON CONFLICT to reset backoff state (not INSERT OR IGNORE)\n- [ ] Discussion sweep deletes documents DIRECTLY (not via dirty queue)\n- [ ] Discussion sweep uses CTE (or Rust-side ID capture) to capture stale IDs before cascading deletes\n- [ ] Documents deleted BEFORE discussions (while source_id still resolves)\n- [ ] ON DELETE CASCADE handles document_labels, document_paths cleanup\n- [ ] documents_embeddings_ad trigger handles embedding cleanup\n- [ ] `cargo build` succeeds\n- [ ] Existing ingestion tests still pass\n\n## Files\n- `src/ingestion/issues.rs` — add mark_dirty_tx calls in upsert loop\n- `src/ingestion/merge_requests.rs` — add mark_dirty_tx calls in upsert loop\n- `src/ingestion/discussions.rs` — add mark_dirty_tx calls in insert loop\n- `src/ingestion/mr_discussions.rs` — add mark_dirty_tx calls + direct document deletion in sweep\n\n## TDD Loop\nRED: Existing tests should still pass (regression); new tests:\n- `test_issue_upsert_marks_dirty` — after issue ingest, dirty_sources has entry\n- `test_mr_upsert_marks_dirty` — after MR ingest, dirty_sources has entry\n- `test_discussion_upsert_marks_dirty` — after discussion ingest, dirty_sources has entry\n- `test_discussion_sweep_deletes_documents` — stale discussion documents deleted directly\n- `test_sweep_cascade_cleans_labels_paths` — ON DELETE CASCADE works\nGREEN: Add mark_dirty_tx calls in all 4 files, implement sweep with CTE\nVERIFY: `cargo test ingestion && cargo build`\n\n## Edge Cases\n- Upsert that doesn't change data: still marks dirty (regenerator hash check handles skip)\n- Transaction rollback: dirty mark also rolled back (atomic, inside same txn)\n- Discussion sweep with zero stale IDs: CTE returns empty, no DELETE executed\n- Large batch of upserts: each mark_dirty_tx is O(1) INSERT with ON CONFLICT\n- Sweep deletes document before discussion: order matters for source_id resolution","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:27:09.540279Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:39:17.241433Z","closed_at":"2026-01-30T17:39:17.241390Z","close_reason":"Added mark_dirty_tx calls in issues.rs, merge_requests.rs, discussions.rs, mr_discussions.rs (2 paths)","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1i2","depends_on_id":"bd-38q","type":"blocks","created_at":"2026-01-30T15:29:35.105551Z","created_by":"tayloreernisse"}]}
+{"id":"bd-1j1","title":"Integration test: full Phase B sync pipeline","description":"## Background\nAfter all Gate 1-2 components are built, we need an integration test proving the full pipeline works end-to-end: sync → enqueue dependent fetches → drain queue → extract refs from state events → parse system notes for refs. Without this, individual unit tests pass but the pipeline may not wire together correctly.\n\n## Approach\nCreate tests/phase_b_integration.rs with a comprehensive test suite:\n\n```rust\n#[tokio::test]\nasync fn test_phase_b_sync_pipeline_integration() {\n    // 1. Create test DB with migrations 001-012\n    // 2. Seed: project, issues, MRs, discussions with system notes\n    // 3. Seed: resource_state_events with source_merge_request_id\n    // 4. Seed: dependent_fetch_queue entries (state_events, label_events)\n    // 5. Run drain_dependent_queue (mocked HTTP → fixture JSON)\n    // 6. Run extract_refs_from_state_events\n    // 7. Run extract_refs_from_system_notes\n    // 8. Assert: entity_references populated with correct source/target/type/method\n    // 9. Assert: no duplicate refs (INSERT OR IGNORE worked)\n    // 10. Assert: unresolved cross-project refs stored correctly\n}\n```\n\nUse wiremock or a trait-based HTTP mock for GitLab API responses. Fixture files in tests/fixtures/phase_b/.\n\n## Acceptance Criteria\n- [ ] Test creates DB, runs all migrations through 012\n- [ ] Test seeds realistic data (issues, MRs, state events, system notes)\n- [ ] Test runs the full pipeline in correct order\n- [ ] Test verifies entity_references from all 3 sources: closes_issues API, state events, system notes\n- [ ] Test verifies deduplication across sources\n- [ ] Test verifies unresolved cross-project references\n- [ ] Test passes with `cargo test phase_b_integration -- --nocapture`\n\n## Files\n- tests/phase_b_integration.rs (new)\n- tests/fixtures/phase_b/state_events.json (new)\n- tests/fixtures/phase_b/label_events.json (new)\n- tests/fixtures/phase_b/system_notes.json (new)\n\n## TDD Loop\nRED: tests/phase_b_integration.rs:\n- `test_full_pipeline_produces_entity_references` - seeds all data, runs full pipeline, asserts entity_references populated from state events + system notes + closes_issues API\n- `test_pipeline_deduplication_across_sources` - same ref discovered by API and system note → single row in entity_references\n- `test_pipeline_unresolved_cross_project_refs` - system note mentioning external project → entity_references row with target_entity_id=NULL and target_iid populated\n- `test_pipeline_empty_queue_succeeds` - no queue entries → pipeline completes with 0 refs, no error\n- `test_pipeline_migrations_001_through_012` - verify all migrations apply cleanly in sequence on fresh DB\n\nSetup: create_test_db helper applying all migrations, seed_phase_b_fixtures() populating issues, MRs, discussions, notes (including system notes with \"closed by !123\" patterns), resource_state_events with source_merge_request fields.\n\nGREEN: Wire pipeline calls in correct order, create fixture JSON files\n\nVERIFY: `cargo test phase_b_integration -- --nocapture`\n\n## Edge Cases\n- Empty queue: pipeline completes successfully with 0 refs\n- All refs duplicate: INSERT OR IGNORE produces 0 new inserts\n- Mixed sources: same ref discovered by API + system note → single entry\n- Migration failure: test should fail clearly if migrations don't apply cleanly","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T22:42:26.355071Z","created_by":"tayloreernisse","updated_at":"2026-02-03T13:42:58.964288Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1j1","depends_on_id":"bd-1ji","type":"blocks","created_at":"2026-02-02T22:43:27.941002Z","created_by":"tayloreernisse"},{"issue_id":"bd-1j1","depends_on_id":"bd-1se","type":"parent-child","created_at":"2026-02-02T22:43:40.577709Z","created_by":"tayloreernisse"},{"issue_id":"bd-1j1","depends_on_id":"bd-3ia","type":"blocks","created_at":"2026-02-02T22:43:28.048311Z","created_by":"tayloreernisse"},{"issue_id":"bd-1j1","depends_on_id":"bd-8t4","type":"blocks","created_at":"2026-02-02T22:43:27.996061Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1je","title":"Implement pending discussion queue","description":"## Background\nThe pending discussion queue tracks discussions that need to be fetched from GitLab. When an issue or MR is updated, its discussions may need re-fetching. This queue is separate from dirty_sources (which tracks entities needing document regeneration) — it tracks entities needing API calls to GitLab. The queue uses the same backoff pattern as dirty_sources for consistency.\n\n## Approach\nCreate `src/ingestion/discussion_queue.rs`:\n\n```rust\nuse crate::core::backoff::compute_next_attempt_at;\n\n/// Noteable type for discussion queue.\n#[derive(Debug, Clone, Copy)]\npub enum NoteableType {\n    Issue,\n    MergeRequest,\n}\n\nimpl NoteableType {\n    pub fn as_str(&self) -> &'static str {\n        match self {\n            Self::Issue => \"Issue\",\n            Self::MergeRequest => \"MergeRequest\",\n        }\n    }\n}\n\npub struct PendingFetch {\n    pub project_id: i64,\n    pub noteable_type: NoteableType,\n    pub noteable_iid: i64,\n    pub attempt_count: i32,\n}\n\n/// Queue a discussion fetch. ON CONFLICT DO UPDATE resets backoff (consistent with dirty_sources).\npub fn queue_discussion_fetch(\n    conn: &Connection,\n    project_id: i64,\n    noteable_type: NoteableType,\n    noteable_iid: i64,\n) -> Result<()>;\n\n/// Get next batch of pending fetches (WHERE next_attempt_at IS NULL OR <= now).\npub fn get_pending_fetches(conn: &Connection, limit: usize) -> Result<Vec<PendingFetch>>;\n\n/// Mark fetch complete (remove from queue).\npub fn complete_fetch(\n    conn: &Connection,\n    project_id: i64,\n    noteable_type: NoteableType,\n    noteable_iid: i64,\n) -> Result<()>;\n\n/// Record fetch error with backoff.\npub fn record_fetch_error(\n    conn: &Connection,\n    project_id: i64,\n    noteable_type: NoteableType,\n    noteable_iid: i64,\n    error: &str,\n) -> Result<()>;\n```\n\n## Acceptance Criteria\n- [ ] queue_discussion_fetch uses ON CONFLICT DO UPDATE (consistent with dirty_sources pattern)\n- [ ] Re-queuing resets: attempt_count=0, next_attempt_at=NULL, last_error=NULL\n- [ ] get_pending_fetches respects next_attempt_at backoff\n- [ ] get_pending_fetches returns entries ordered by queued_at ASC\n- [ ] complete_fetch removes entry from queue\n- [ ] record_fetch_error increments attempt_count, computes next_attempt_at via shared backoff\n- [ ] NoteableType.as_str() returns \"Issue\" or \"MergeRequest\" (matches DB CHECK constraint)\n- [ ] `cargo test discussion_queue` passes\n\n## Files\n- `src/ingestion/discussion_queue.rs` — new file\n- `src/ingestion/mod.rs` — add `pub mod discussion_queue;`\n\n## TDD Loop\nRED: Tests in `#[cfg(test)] mod tests`:\n- `test_queue_and_get` — queue entry, get returns it\n- `test_requeue_resets_backoff` — queue, error, re-queue -> attempt_count=0\n- `test_backoff_respected` — entry with future next_attempt_at not returned\n- `test_complete_removes` — complete_fetch removes entry\n- `test_error_increments_attempts` — error -> attempt_count=1, next_attempt_at set\nGREEN: Implement all functions\nVERIFY: `cargo test discussion_queue`\n\n## Edge Cases\n- Queue same (project_id, noteable_type, noteable_iid) twice: ON CONFLICT resets state\n- NoteableType must match DB CHECK constraint exactly (\"Issue\", \"MergeRequest\" — capitalized)\n- Empty queue: get_pending_fetches returns empty Vec","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:27:09.505548Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:31:35.496454Z","closed_at":"2026-01-30T17:31:35.496405Z","close_reason":"Implemented discussion_queue with queue/get/complete/record_error + 6 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1je","depends_on_id":"bd-hrs","type":"blocks","created_at":"2026-01-30T15:29:35.034753Z","created_by":"tayloreernisse"},{"issue_id":"bd-1je","depends_on_id":"bd-mem","type":"blocks","created_at":"2026-01-30T15:29:35.071573Z","created_by":"tayloreernisse"}]}
+{"id":"bd-1ji","title":"Parse system notes for cross-reference patterns","description":"## Background\nSystem notes contain cross-reference patterns like 'mentioned in !{iid}', 'closed by !{iid}', etc. This is best-effort, English-only extraction that supplements the structured API data from bd-3ia and bd-8t4. Runs as a local post-processing step (no API calls).\n\n## Approach\nCreate src/core/note_parser.rs:\n\n```rust\nuse regex::Regex;\nuse lazy_static::lazy_static;\n\n/// A parsed cross-reference from a system note.\npub struct ParsedCrossRef {\n    pub reference_type: String,    // \"mentioned\" | \"closes\"\n    pub target_entity_type: String, // \"issue\" | \"merge_request\"  \n    pub target_iid: i64,\n    pub target_project_path: Option<String>,  // None = same project\n}\n\nlazy_static! {\n    static ref MENTIONED_RE: Regex = Regex::new(\n        r\"mentioned in (?:(?P<project>[\\w\\-]+/[\\w\\-]+))?(?P<sigil>[#!])(?P<iid>\\d+)\"\n    ).unwrap();\n    static ref CLOSED_BY_RE: Regex = Regex::new(\n        r\"closed by (?:(?P<project>[\\w\\-]+/[\\w\\-]+))?(?P<sigil>[#!])(?P<iid>\\d+)\"\n    ).unwrap();\n}\n\n/// Parse a system note body for cross-references.\npub fn parse_cross_refs(body: &str) -> Vec<ParsedCrossRef>\n\n/// Extract cross-references from all system notes and insert into entity_references.\n/// Queries notes WHERE is_system = 1, parses body text, resolves to entity_references.\npub fn extract_refs_from_system_notes(\n    conn: &Connection,\n    project_id: i64,\n) -> Result<ExtractResult>\n\npub struct ExtractResult {\n    pub inserted: usize,\n    pub skipped_unresolvable: usize,\n    pub parse_failures: usize,   // logged at debug level\n}\n```\n\nSigil mapping: `#` = issue, `!` = merge_request\n\nResolution logic:\n1. If target_project_path is None (same project): look up entity by iid in local DB → set target_entity_id\n2. If target_project_path is Some: check if project is synced locally\n   - If yes: resolve to local entity id\n   - If no: store as unresolved (target_entity_id=NULL, target_project_path=path, target_entity_iid=iid)\n\nInsert with source_method='system_note_parse', INSERT OR IGNORE for dedup.\n\nCall after drain_dependent_queue and extract_refs_from_state_events in the sync pipeline.\n\n## Acceptance Criteria\n- [ ] 'mentioned in !123' → mentioned ref, target=MR iid 123\n- [ ] 'mentioned in #456' → mentioned ref, target=issue iid 456\n- [ ] 'mentioned in group/project!789' → cross-project mentioned ref\n- [ ] 'closed by !123' → closes ref\n- [ ] Cross-project refs stored as unresolved when target project not synced\n- [ ] source_method = 'system_note_parse'\n- [ ] Parse failures logged at debug level (not errors)\n- [ ] Idempotent (INSERT OR IGNORE)\n- [ ] Only processes is_system=1 notes\n\n## Files\n- src/core/note_parser.rs (new)\n- src/core/mod.rs (add `pub mod note_parser;`)\n- src/cli/commands/sync.rs (call after other ref extraction steps)\n\n## TDD Loop\nRED: tests/note_parser_tests.rs:\n- `test_parse_mentioned_in_mr` - \"mentioned in !567\" → ParsedCrossRef { mentioned, merge_request, 567 }\n- `test_parse_mentioned_in_issue` - \"mentioned in #234\" → ParsedCrossRef { mentioned, issue, 234 }\n- `test_parse_mentioned_cross_project` - \"mentioned in group/repo!789\" → with project path\n- `test_parse_closed_by_mr` - \"closed by !567\" → ParsedCrossRef { closes, merge_request, 567 }\n- `test_parse_multiple_refs` - note with two mentions → two refs\n- `test_parse_no_refs` - \"Updated the description\" → empty vec\n- `test_extract_refs_from_system_notes_integration` - seed DB with system notes, verify entity_references created\n\nGREEN: Implement regex patterns and extraction logic\n\nVERIFY: `cargo test note_parser -- --nocapture`\n\n## Edge Cases\n- Non-English GitLab instances: \"ajouté l'étiquette ~bug\" won't match — this is accepted limitation, logged at debug\n- Multi-level group paths: \"mentioned in top/sub/project#123\" — regex needs to handle arbitrary depth ([\\w\\-]+(?:/[\\w\\-]+)+)\n- Note body may contain markdown links that look like refs: \"[#123](url)\" — the regex should handle this correctly since the prefix \"mentioned in\" is required\n- Same ref mentioned multiple times in same note — dedup via INSERT OR IGNORE\n- Note may reference itself (e.g., system note on issue #123 says \"mentioned in #123\") — technically valid, store it","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T21:32:33.663304Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:41:50.672968Z","compaction_level":0,"original_size":0,"labels":["gate-2","parsing","phase-b"],"dependencies":[{"issue_id":"bd-1ji","depends_on_id":"bd-1se","type":"parent-child","created_at":"2026-02-02T21:32:33.665218Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ji","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T22:41:50.672947Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1k1","title":"Implement FTS5 search function and query sanitization","description":"## Background\nFTS5 search is the core lexical retrieval engine. It wraps SQLite's FTS5 with safe query parsing that prevents user input from causing SQL syntax errors, while preserving useful features like prefix search for type-ahead. The search function returns ranked results with BM25 scores and contextual snippets. This module is the Gate A search backbone and also provides fallback search when Ollama is unavailable in Gate B.\n\n## Approach\nCreate `src/search/` module with `mod.rs` and `fts.rs` per PRD Section 3.1-3.2.\n\n**src/search/mod.rs:**\n```rust\nmod fts;\nmod filters;\n// Later beads add: mod vector; mod hybrid; mod rrf;\npub use fts::{search_fts, to_fts_query, FtsResult, FtsQueryMode, generate_fallback_snippet, get_result_snippet};\n```\n\n**src/search/fts.rs — key functions:**\n\n1. `to_fts_query(raw: &str, mode: FtsQueryMode) -> String`\n   - Safe mode: wrap each token in quotes, escape internal quotes, preserve trailing * on alphanumeric tokens\n   - Raw mode: pass through unchanged\n\n2. `search_fts(conn: &Connection, query: &str, limit: usize, mode: FtsQueryMode) -> Result<Vec<FtsResult>>`\n   - Uses `bm25(documents_fts)` for ranking\n   - Uses `snippet(documents_fts, 1, '<mark>', '</mark>', '...', 64)` for context\n   - Column index 1 = content_text (0=title)\n\n3. `generate_fallback_snippet(content_text: &str, max_chars: usize) -> String`\n   - For semantic-only results without FTS snippets\n   - Uses `truncate_utf8()` for safe byte boundaries\n\n4. `truncate_utf8(s: &str, max_bytes: usize) -> &str`\n   - Walks backward from max_bytes to find nearest char boundary\n\n5. `get_result_snippet(fts_snippet: Option<&str>, content_text: &str) -> String`\n   - Prefers FTS snippet, falls back to truncated content\n\nUpdate `src/lib.rs`: add `pub mod search;`\n\n## Acceptance Criteria\n- [ ] Porter stemming works: search \"searching\" matches document containing \"search\"\n- [ ] Prefix search works: `auth*` matches \"authentication\"\n- [ ] Empty query returns empty Vec (no error)\n- [ ] Special characters don't cause FTS5 errors: `-`, `\"`, `:`, `*`\n- [ ] Query `\"-DWITH_SSL\"` returns results (dash not treated as NOT operator)\n- [ ] Query `C++` returns results (special chars preserved in quotes)\n- [ ] Safe mode preserves trailing `*` on alphanumeric tokens: `auth*` -> `\"auth\"*`\n- [ ] Raw mode passes query unchanged\n- [ ] BM25 scores returned (lower = better match)\n- [ ] Snippets contain `<mark>` tags around matches\n- [ ] `generate_fallback_snippet` truncates at word boundary, appends \"...\"\n- [ ] `truncate_utf8` never panics on multi-byte codepoints\n- [ ] `cargo test fts` passes\n\n## Files\n- `src/search/mod.rs` — new file (module root)\n- `src/search/fts.rs` — new file (FTS5 search + query sanitization)\n- `src/lib.rs` — add `pub mod search;`\n\n## TDD Loop\nRED: Tests in `fts.rs` `#[cfg(test)] mod tests`:\n- `test_safe_query_basic` — \"auth error\" -> `\"auth\" \"error\"`\n- `test_safe_query_prefix` — \"auth*\" -> `\"auth\"*`\n- `test_safe_query_special_chars` — \"C++\" -> `\"C++\"`\n- `test_safe_query_dash` — \"-DWITH_SSL\" -> `\"-DWITH_SSL\"`\n- `test_safe_query_quotes` — `he said \"hello\"` -> escaped\n- `test_raw_mode_passthrough` — raw query unchanged\n- `test_empty_query` — returns empty vec\n- `test_truncate_utf8_emoji` — truncate mid-emoji walks back\n- `test_fallback_snippet_word_boundary` — truncates at space\nGREEN: Implement to_fts_query, search_fts, helpers\nVERIFY: `cargo test fts`\n\n## Edge Cases\n- Query with only whitespace: treated as empty, returns empty\n- Query with only special characters: quoted, may return no results (not an error)\n- Very long query (1000+ chars): works but may be slow (no explicit limit)\n- FTS5 snippet returns empty string: fallback to truncated content_text\n- Non-alphanumeric prefix: `C++*` — NOT treated as prefix (special chars present)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:13.005179Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:23:35.204290Z","closed_at":"2026-01-30T17:23:35.204106Z","close_reason":"Completed: to_fts_query (safe/raw modes), search_fts with BM25+snippets, generate_fallback_snippet, get_result_snippet, truncate_utf8 reuse, 13 tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1k1","depends_on_id":"bd-221","type":"blocks","created_at":"2026-01-30T15:29:24.374108Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1kh","title":"[CP0] Raw payload handling - compression and deduplication","description":"## Background\n\nRaw payload storage allows replaying API responses for debugging and audit. Compression reduces storage for large payloads. SHA-256 deduplication prevents storing identical payloads multiple times (important for frequently polled resources that haven't changed).\n\nReference: docs/prd/checkpoint-0.md section \"Raw Payload Handling\"\n\n## Approach\n\n**src/core/payloads.ts:**\n```typescript\nimport { createHash } from 'node:crypto';\nimport { gzipSync, gunzipSync } from 'node:zlib';\nimport Database from 'better-sqlite3';\nimport { nowMs } from './time';\n\ninterface StorePayloadOptions {\n  projectId: number | null;\n  resourceType: string;  // 'project' | 'issue' | 'mr' | 'note' | 'discussion'\n  gitlabId: string;      // TEXT because discussion IDs are strings\n  payload: unknown;      // JSON-serializable object\n  compress: boolean;     // from config.storage.compressRawPayloads\n}\n\nexport function storePayload(db: Database.Database, options: StorePayloadOptions): number | null {\n  // 1. JSON.stringify the payload\n  // 2. SHA-256 hash the JSON bytes\n  // 3. Check for duplicate by (project_id, resource_type, gitlab_id, payload_hash)\n  // 4. If duplicate, return existing ID\n  // 5. If compress=true, gzip the JSON bytes\n  // 6. INSERT with content_encoding='gzip' or 'identity'\n  // 7. Return lastInsertRowid\n}\n\nexport function readPayload(db: Database.Database, id: number): unknown {\n  // 1. SELECT content_encoding, payload FROM raw_payloads WHERE id = ?\n  // 2. If gzip, decompress\n  // 3. JSON.parse and return\n}\n```\n\n## Acceptance Criteria\n\n- [ ] storePayload() with compress=true stores gzip-encoded payload\n- [ ] storePayload() with compress=false stores identity-encoded payload\n- [ ] Duplicate payload (same hash) returns existing row ID, not new row\n- [ ] readPayload() correctly decompresses gzip payloads\n- [ ] readPayload() returns null for non-existent ID\n- [ ] SHA-256 hash computed from pre-compression JSON bytes\n- [ ] Large payloads (100KB+) compress to ~10-20% of original size\n\n## Files\n\nCREATE:\n- src/core/payloads.ts\n- tests/unit/payloads.test.ts\n\n## TDD Loop\n\nRED:\n```typescript\n// tests/unit/payloads.test.ts\ndescribe('Payload Storage', () => {\n  describe('storePayload', () => {\n    it('stores uncompressed payload with identity encoding')\n    it('stores compressed payload with gzip encoding')\n    it('deduplicates identical payloads by hash')\n    it('stores different payloads for same gitlab_id')\n  })\n\n  describe('readPayload', () => {\n    it('reads uncompressed payload')\n    it('reads and decompresses gzip payload')\n    it('returns null for non-existent id')\n  })\n})\n```\n\nGREEN: Implement storePayload() and readPayload()\n\nVERIFY: `npm run test -- tests/unit/payloads.test.ts`\n\n## Edge Cases\n\n- gitlabId is TEXT not INTEGER - discussion IDs are UUIDs\n- Compression ratio varies - some JSON compresses better than others\n- null projectId valid for global resources (like user profile)\n- Hash collision extremely unlikely with SHA-256 but unique index enforces","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:50.189494Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:19:12.854771Z","closed_at":"2026-01-25T03:19:12.854372Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1kh","depends_on_id":"bd-3ng","type":"blocks","created_at":"2026-01-24T16:13:09.055338Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1l1","title":"[CP0] GitLab API client with rate limiting","description":"## Background\n\nThe GitLab client handles all API communication with rate limiting to avoid 429 errors. Uses native fetch (Node 18+). Rate limiter adds jitter to prevent thundering herd. All errors are typed for clean error handling in CLI commands.\n\nReference: docs/prd/checkpoint-0.md section \"GitLab Client\"\n\n## Approach\n\n**src/gitlab/client.ts:**\n```typescript\nexport class GitLabClient {\n  private baseUrl: string;\n  private token: string;\n  private rateLimiter: RateLimiter;\n\n  constructor(options: { baseUrl: string; token: string; requestsPerSecond?: number }) {\n    this.baseUrl = options.baseUrl.replace(/\\/$/, '');\n    this.token = options.token;\n    this.rateLimiter = new RateLimiter(options.requestsPerSecond ?? 10);\n  }\n\n  async getCurrentUser(): Promise<GitLabUser>\n  async getProject(pathWithNamespace: string): Promise<GitLabProject>\n  private async request<T>(path: string, options?: RequestInit): Promise<T>\n}\n\nclass RateLimiter {\n  private lastRequest = 0;\n  private minInterval: number;\n\n  constructor(requestsPerSecond: number) {\n    this.minInterval = 1000 / requestsPerSecond;\n  }\n\n  async acquire(): Promise<void> {\n    // Wait if too soon since last request\n    // Add 0-50ms jitter\n  }\n}\n```\n\n**src/gitlab/types.ts:**\n```typescript\nexport interface GitLabUser {\n  id: number;\n  username: string;\n  name: string;\n}\n\nexport interface GitLabProject {\n  id: number;\n  path_with_namespace: string;\n  default_branch: string;\n  web_url: string;\n  created_at: string;\n  updated_at: string;\n}\n```\n\n**Integration tests with MSW (Mock Service Worker):**\nSet up MSW handlers that mock GitLab API responses for /api/v4/user and /api/v4/projects/:path\n\n## Acceptance Criteria\n\n- [ ] getCurrentUser() returns GitLabUser with id, username, name\n- [ ] getProject(\"group/project\") URL-encodes path correctly\n- [ ] 401 response throws GitLabAuthError\n- [ ] 404 response throws GitLabNotFoundError\n- [ ] 429 response throws GitLabRateLimitError with retryAfter from header\n- [ ] Network failure throws GitLabNetworkError\n- [ ] Rate limiter enforces minimum interval between requests\n- [ ] Rate limiter adds random jitter (0-50ms)\n- [ ] tests/integration/gitlab-client.test.ts passes (6 tests)\n\n## Files\n\nCREATE:\n- src/gitlab/client.ts\n- src/gitlab/types.ts\n- tests/integration/gitlab-client.test.ts\n- tests/fixtures/mock-responses/gitlab-user.json\n- tests/fixtures/mock-responses/gitlab-project.json\n\n## TDD Loop\n\nRED:\n```typescript\n// tests/integration/gitlab-client.test.ts\ndescribe('GitLab Client', () => {\n  it('authenticates with valid PAT')\n  it('returns 401 for invalid PAT')\n  it('fetches project by path')\n  it('handles rate limiting (429) with Retry-After')\n  it('respects rate limit (requests per second)')\n  it('adds jitter to rate limiting')\n})\n```\n\nGREEN: Implement client.ts and types.ts\n\nVERIFY: `npm run test -- tests/integration/gitlab-client.test.ts`\n\n## Edge Cases\n\n- Path with special characters (spaces, slashes) must be URL-encoded\n- Retry-After header may be missing - default to 60s\n- Network timeout should be handled (use AbortController)\n- Rate limiter jitter prevents multiple clients syncing in lockstep\n- baseUrl trailing slash should be stripped","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:49.842981Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:06:39.520300Z","closed_at":"2026-01-25T03:06:39.520131Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1l1","depends_on_id":"bd-gg1","type":"blocks","created_at":"2026-01-24T16:13:08.713272Z","created_by":"tayloreernisse"}]}
+{"id":"bd-1m8","title":"Extend 'lore stats --check' for event table integrity and queue health","description":"## Background\nThe existing stats --check command validates data integrity. Need to extend it for event tables (referential integrity) and dependent job queue health (stuck locks, retryable jobs). This provides operators and agents a way to detect data quality issues after sync.\n\n## Approach\nExtend src/cli/commands/stats.rs check mode:\n\n**New checks:**\n\n1. Event FK integrity:\n```sql\n-- Orphaned state events (issue_id points to non-existent issue)\nSELECT COUNT(*) FROM resource_state_events rse\nWHERE rse.issue_id IS NOT NULL\n  AND NOT EXISTS (SELECT 1 FROM issues i WHERE i.id = rse.issue_id);\n-- (repeat for merge_request_id, and for label + milestone event tables)\n```\n\n2. Queue health:\n```sql\n-- Pending jobs by type\nSELECT job_type, COUNT(*) FROM pending_dependent_fetches GROUP BY job_type;\n-- Stuck locks (locked_at older than 5 minutes)\nSELECT COUNT(*) FROM pending_dependent_fetches WHERE locked_at IS NOT NULL AND locked_at < ?;\n-- Retryable jobs (attempts > 0, not locked)\nSELECT COUNT(*) FROM pending_dependent_fetches WHERE attempts > 0 AND locked_at IS NULL;\n-- Max attempts (jobs that may be permanently failing)\nSELECT job_type, MAX(attempts) FROM pending_dependent_fetches GROUP BY job_type;\n```\n\n3. Human output per check: PASS / WARN / FAIL with counts\n```\nEvent FK integrity:     PASS (0 orphaned events)\nQueue health:           WARN (3 stuck locks, 12 retryable jobs)\n```\n\n4. Robot JSON: structured health report\n```json\n{\n  \"event_integrity\": {\n    \"status\": \"pass\",\n    \"orphaned_state_events\": 0,\n    \"orphaned_label_events\": 0,\n    \"orphaned_milestone_events\": 0\n  },\n  \"queue_health\": {\n    \"status\": \"warn\",\n    \"pending_by_type\": {\"resource_events\": 5, \"mr_closes_issues\": 2},\n    \"stuck_locks\": 3,\n    \"retryable_jobs\": 12,\n    \"max_attempts_by_type\": {\"resource_events\": 5}\n  }\n}\n```\n\n## Acceptance Criteria\n- [ ] Detects orphaned events (FK target missing)\n- [ ] Detects stuck locks (locked_at older than threshold)\n- [ ] Reports retryable job count and max attempts\n- [ ] Human output shows PASS/WARN/FAIL per check\n- [ ] Robot JSON matches structured schema\n- [ ] Graceful when event/queue tables don't exist\n\n## Files\n- src/cli/commands/stats.rs (extend check mode)\n\n## TDD Loop\nRED: tests/stats_check_tests.rs:\n- `test_stats_check_events_pass` - clean data, verify PASS\n- `test_stats_check_events_orphaned` - delete an issue with events remaining, verify FAIL count\n- `test_stats_check_queue_stuck_locks` - set old locked_at, verify WARN\n- `test_stats_check_queue_retryable` - fail some jobs, verify retryable count\n\nGREEN: Add the check queries and formatting\n\nVERIFY: `cargo test stats_check -- --nocapture`\n\n## Edge Cases\n- FK with CASCADE should prevent orphaned events in normal operation — but manual DB edits or bugs could cause them\n- Tables may not exist if migration 011 not applied — check table existence before querying\n- Empty queue is PASS (not WARN for \"no jobs found\")\n- Distinguish between \"0 stuck locks\" (good) and \"queue table doesn't exist\" (skip check)","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T21:31:57.422916Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:43:40.182992Z","compaction_level":0,"original_size":0,"labels":["cli","gate-1","phase-b"],"dependencies":[{"issue_id":"bd-1m8","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:57.424103Z","created_by":"tayloreernisse"},{"issue_id":"bd-1m8","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T21:32:06.350605Z","created_by":"tayloreernisse"},{"issue_id":"bd-1m8","depends_on_id":"bd-tir","type":"blocks","created_at":"2026-02-02T21:32:06.391042Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1mf","title":"[CP1] gi sync-status enhancement","description":"Enhance sync-status from CP0 stub to show issue cursors.\n\nOutput:\n- Last run timestamp and duration\n- Cursor positions per project (issues resource_type)\n- Entity counts (issues, discussions, notes)\n\nFiles: src/cli/commands/sync-status.ts (update existing)\nDone when: Shows cursor positions and counts after ingestion","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T15:20:36.449088Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.157235Z","deleted_at":"2026-01-25T15:21:35.157232Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-1n5","title":"[CP1] gi ingest --type=issues command","description":"CLI command to orchestrate issue ingestion.\n\nImplementation:\n1. Acquire app lock with heartbeat\n2. Create sync_run record (status='running')\n3. For each configured project:\n   - Call ingestIssues()\n   - For each ingested issue, call ingestIssueDiscussions()\n   - Show progress (spinner or progress bar)\n4. Update sync_run (status='succeeded', metrics_json)\n5. Release lock\n\nFlags:\n- --type=issues (required)\n- --project=PATH (optional, filter to single project)\n- --force (override stale lock)\n\nOutput: Progress bar, then summary with counts\n\nFiles: src/cli/commands/ingest.ts\nTests: tests/integration/sync-runs.test.ts\nDone when: Full issue + discussion ingestion works end-to-end","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:20:05.114751Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.153598Z","deleted_at":"2026-01-25T15:21:35.153595Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-1nf","title":"Register 'lore timeline' command with all flags","description":"## Background\nWire the timeline command into clap CLI with all flags. This is the integration point that connects the query engine (seed → expand → collect) with the output renderers (human/robot).\n\n## Approach\n1. Add to src/cli/mod.rs Commands enum:\n```rust\n/// Show chronological decision timeline for a keyword\n#[command(name = \"timeline\")]\nTimeline(TimelineArgs),\n```\n\n```rust\n#[derive(clap::Args)]\npub struct TimelineArgs {\n    /// Keyword or phrase to search for\n    #[arg(required = true)]\n    pub query: String,\n\n    /// Scope to a specific project\n    #[arg(short = 'p', long = \"project\")]\n    pub project: Option<String>,\n\n    /// Only include events after this date (Nd, Nw, Nm, or YYYY-MM-DD)\n    #[arg(long)]\n    pub since: Option<String>,\n\n    /// Cross-reference expansion depth (0 = disabled)\n    #[arg(long, default_value = \"1\")]\n    pub depth: usize,\n\n    /// Also follow 'mentioned' edges during expansion\n    #[arg(long = \"expand-mentions\")]\n    pub expand_mentions: bool,\n\n    /// Maximum number of events to return\n    #[arg(short = 'n', long = \"limit\", default_value = \"100\")]\n    pub limit: usize,\n}\n```\n\n2. Add handler in src/main.rs:\n```rust\nCommands::Timeline(args) => handle_timeline(cli.config.as_deref(), args, robot_mode).await,\n```\n\n3. Implement handle_timeline:\n```rust\nasync fn handle_timeline(config_override: Option<&str>, args: TimelineArgs, robot_mode: bool) -> Result<...> {\n    let config = Config::load(config_override)?;\n    let db_path = get_db_path(config.storage.db_path.as_deref());\n    let conn = create_connection(&db_path)?;\n    \n    let project_id = args.project.as_ref()\n        .map(|p| resolve_project(&conn, p))\n        .transpose()?;\n    let since = args.since.as_ref()\n        .map(|s| parse_since(s))   // reuse existing time parsing from core/time.rs\n        .transpose()?;\n    \n    // Pipeline: seed → expand → collect\n    let seed = seed_from_fts(&conn, &args.query, project_id, since, args.limit)?;\n    let expansion = expand_via_references(&conn, &seed.entities, args.depth, args.expand_mentions, 100)?;\n    let events = collect_and_interleave_events(&conn, &seed.entities, &expansion.expanded, &seed.evidence_notes, Some(args.limit))?;\n    \n    let result = TimelineResult { query: args.query, events, seed_entities: ..., expanded_entities: ..., unresolved_references: ... };\n    \n    if robot_mode {\n        print_timeline_json(&result, args.depth, args.expand_mentions);\n    } else {\n        print_timeline(&result, use_color);\n    }\n}\n```\n\n4. Parse --since: reuse existing parse_since/parse_duration from src/core/time.rs. Support:\n   - Relative: 7d, 2w, 6m (days, weeks, months)\n   - Absolute: YYYY-MM-DD\n\n5. Export from src/cli/commands/mod.rs: `pub mod timeline;`\n\n## Acceptance Criteria\n- [ ] `lore timeline \"auth\"` works with all defaults\n- [ ] `lore timeline \"auth\" -p group/repo` scopes to project\n- [ ] `lore timeline \"auth\" --since 7d` filters by date\n- [ ] `lore timeline \"auth\" --depth 0` disables expansion\n- [ ] `lore timeline \"auth\" --depth 2` expands to 2 hops\n- [ ] `lore timeline \"auth\" --expand-mentions` includes mentioned edges\n- [ ] `lore timeline \"auth\" -n 50` limits events\n- [ ] `lore -J timeline \"auth\"` outputs robot JSON\n- [ ] --since supports relative (7d, 2w, 6m) and absolute (YYYY-MM-DD)\n- [ ] Help text is descriptive for all flags\n\n## Files\n- src/cli/mod.rs (add Timeline variant + TimelineArgs)\n- src/cli/commands/mod.rs (add `pub mod timeline;`)\n- src/main.rs (add handler)\n\n## TDD Loop\nRED: tests/cli_tests.rs (or integration):\n- `test_timeline_cli_parse_defaults` - verify default depth=1, limit=100\n- `test_timeline_cli_parse_all_flags` - verify all flags parsed\n- `test_timeline_cli_since_relative` - verify 7d parsed to ms epoch\n- `test_timeline_cli_since_absolute` - verify 2024-01-01 parsed\n\nGREEN: Add CLI definition, handler, and wiring\n\nVERIFY: `cargo build && cargo test cli -- --nocapture`\n\n## Edge Cases\n- Empty query string: clap requires it, but could be whitespace-only — validate non-empty after trim\n- --since with invalid format: return user-friendly error with examples\n- --depth > 5 could be expensive — consider warning or cap\n- Project resolution may be ambiguous — existing resolve_project handles this with exit code 18","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:28.422082Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:47:42.363175Z","compaction_level":0,"original_size":0,"labels":["cli","gate-3","phase-b"],"dependencies":[{"issue_id":"bd-1nf","depends_on_id":"bd-2f2","type":"blocks","created_at":"2026-02-02T21:33:37.746192Z","created_by":"tayloreernisse"},{"issue_id":"bd-1nf","depends_on_id":"bd-dty","type":"blocks","created_at":"2026-02-02T21:33:37.788079Z","created_by":"tayloreernisse"},{"issue_id":"bd-1nf","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:28.423399Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1np","title":"[CP1] GitLab types for issues, discussions, notes","description":"## Background\n\nGitLab types define the Rust structs for deserializing GitLab API responses. These types are the foundation for all ingestion work - issues, discussions, and notes must be correctly typed for serde to parse them.\n\n## Approach\n\nAdd types to `src/gitlab/types.rs` with serde derives:\n\n### GitLabIssue\n\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabIssue {\n    pub id: i64,                      // GitLab global ID\n    pub iid: i64,                     // Project-scoped issue number\n    pub project_id: i64,\n    pub title: String,\n    pub description: Option<String>,\n    pub state: String,                // \"opened\" | \"closed\"\n    pub created_at: String,           // ISO 8601\n    pub updated_at: String,           // ISO 8601\n    pub closed_at: Option<String>,\n    pub author: GitLabAuthor,\n    pub labels: Vec<String>,          // Array of label names (CP1 canonical)\n    pub web_url: String,\n}\n```\n\nNOTE: `labels_details` intentionally NOT modeled - varies across GitLab versions.\n\n### GitLabAuthor\n\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabAuthor {\n    pub id: i64,\n    pub username: String,\n    pub name: String,\n}\n```\n\n### GitLabDiscussion\n\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabDiscussion {\n    pub id: String,                   // String ID like \"6a9c1750b37d...\"\n    pub individual_note: bool,        // true = standalone comment\n    pub notes: Vec<GitLabNote>,\n}\n```\n\n### GitLabNote\n\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabNote {\n    pub id: i64,\n    #[serde(rename = \"type\")]\n    pub note_type: Option<String>,    // \"DiscussionNote\" | \"DiffNote\" | null\n    pub body: String,\n    pub author: GitLabAuthor,\n    pub created_at: String,           // ISO 8601\n    pub updated_at: String,           // ISO 8601\n    pub system: bool,                 // true for system-generated notes\n    #[serde(default)]\n    pub resolvable: bool,\n    #[serde(default)]\n    pub resolved: bool,\n    pub resolved_by: Option<GitLabAuthor>,\n    pub resolved_at: Option<String>,\n    pub position: Option<GitLabNotePosition>,\n}\n```\n\n### GitLabNotePosition\n\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabNotePosition {\n    pub old_path: Option<String>,\n    pub new_path: Option<String>,\n    pub old_line: Option<i32>,\n    pub new_line: Option<i32>,\n}\n```\n\n## Acceptance Criteria\n\n- [ ] GitLabIssue deserializes from API response JSON\n- [ ] GitLabAuthor embedded correctly in issue and note\n- [ ] GitLabDiscussion with notes array deserializes\n- [ ] GitLabNote handles null note_type (use Option<String>)\n- [ ] GitLabNote uses #[serde(rename = \"type\")] for reserved keyword\n- [ ] resolvable/resolved default to false via #[serde(default)]\n- [ ] All timestamp fields are String (ISO 8601 parsed elsewhere)\n\n## Files\n\n- src/gitlab/types.rs (edit - add types)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/gitlab_types_tests.rs\n#[test] fn deserializes_gitlab_issue_from_json()\n#[test] fn deserializes_gitlab_discussion_from_json()\n#[test] fn handles_null_note_type()\n#[test] fn handles_missing_resolvable_field()\n#[test] fn deserializes_labels_as_string_array()\n```\n\nGREEN: Add type definitions with serde attributes\n\nVERIFY: `cargo test gitlab_types`\n\n## Edge Cases\n\n- note_type can be null, \"DiscussionNote\", or \"DiffNote\"\n- labels array can be empty\n- description can be null\n- resolved_by/resolved_at can be null\n- position is only present for DiffNotes","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.150472Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:17:08.842965Z","closed_at":"2026-01-25T22:17:08.842895Z","close_reason":"Implemented GitLabAuthor, GitLabIssue, GitLabDiscussion, GitLabNote, GitLabNotePosition types with 10 passing tests","compaction_level":0,"original_size":0}
+{"id":"bd-1oo","title":"Write migration 012: mr_file_changes + commit SHA columns","description":"## Background\nNeed to track which files each MR touched, plus commit SHAs for git linking. This is migration 012 (since 011 is resource events + entity_references + dependent fetch queue).\n\n## Approach\nCreate migrations/012_file_changes.sql with the exact schema from spec §4.1:\n\n```sql\n-- Files changed by each merge request\nCREATE TABLE mr_file_changes (\n  id INTEGER PRIMARY KEY,\n  merge_request_id INTEGER NOT NULL REFERENCES merge_requests(id) ON DELETE CASCADE,\n  project_id INTEGER NOT NULL REFERENCES projects(id) ON DELETE CASCADE,\n  old_path TEXT,\n  new_path TEXT NOT NULL,\n  change_type TEXT NOT NULL CHECK (change_type IN ('added', 'modified', 'deleted', 'renamed')),\n  UNIQUE(merge_request_id, new_path)\n);\n\nCREATE INDEX idx_mr_files_new_path ON mr_file_changes(new_path);\nCREATE INDEX idx_mr_files_old_path ON mr_file_changes(old_path) WHERE old_path IS NOT NULL;\nCREATE INDEX idx_mr_files_mr ON mr_file_changes(merge_request_id);\n\n-- Add commit SHAs to merge_requests (cherry-picked from Phase A)\nALTER TABLE merge_requests ADD COLUMN merge_commit_sha TEXT;\nALTER TABLE merge_requests ADD COLUMN squash_commit_sha TEXT;\n\nINSERT INTO schema_version (version, applied_at, description)\nVALUES (12, strftime('%s', 'now') * 1000, 'MR file changes and commit SHA columns');\n```\n\nRegister in src/core/db.rs MIGRATIONS array:\n```rust\n(\"012\", include_str!(\"../../migrations/012_file_changes.sql\")),\n```\n\n## Acceptance Criteria\n- [ ] migrations/012_file_changes.sql exists with mr_file_changes table + indexes\n- [ ] merge_requests table has merge_commit_sha and squash_commit_sha columns\n- [ ] src/core/db.rs MIGRATIONS array includes (\"012\", ...)\n- [ ] Migration applies cleanly after 011\n- [ ] `cargo test migration` passes\n- [ ] UNIQUE(merge_request_id, new_path) enforced\n- [ ] change_type CHECK constraint enforced\n\n## Files\n- migrations/012_file_changes.sql (new)\n- src/core/db.rs (add to MIGRATIONS array)\n\n## TDD Loop\nRED: tests/migration_tests.rs:\n- `test_migration_012_creates_mr_file_changes` - verify table exists\n- `test_migration_012_adds_commit_sha_columns` - verify columns on merge_requests\n- `test_migration_012_change_type_constraint` - verify CHECK rejects invalid values\n- `test_migration_012_unique_constraint` - verify duplicate (mr_id, new_path) rejected\n\nGREEN: Write the migration SQL + register in db.rs\n\nVERIFY: `cargo test migration_tests -- --nocapture`\n\n## Edge Cases\n- ALTER TABLE on merge_requests with existing data — SQLite handles this gracefully (new columns default to NULL)\n- old_path can be NULL (new files) or same as new_path (modified files) — only populated for renames\n- If migration 011 hasn't been applied, 012 should still fail cleanly (schema_version check in runner)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:34:08.837816Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:48:00.372836Z","compaction_level":0,"original_size":0,"labels":["gate-4","phase-b","schema"],"dependencies":[{"issue_id":"bd-1oo","depends_on_id":"bd-14q","type":"parent-child","created_at":"2026-02-02T21:34:08.843541Z","created_by":"tayloreernisse"},{"issue_id":"bd-1oo","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T21:34:16.505965Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1qf","title":"[CP1] Discussion and note transformers","description":"## Background\n\nDiscussion and note transformers convert GitLab API discussion responses into our normalized schema. They compute derived fields like `first_note_at`, `last_note_at`, resolvable/resolved status, and note positions. These are pure functions with no I/O.\n\n## Approach\n\nCreate transformer module with:\n\n### Structs\n\n```rust\n// src/gitlab/transformers/discussion.rs\n\npub struct NormalizedDiscussion {\n    pub gitlab_discussion_id: String,\n    pub project_id: i64,\n    pub issue_id: i64,\n    pub noteable_type: String,        // \"Issue\"\n    pub individual_note: bool,\n    pub first_note_at: Option<i64>,   // min(note.created_at) in ms epoch\n    pub last_note_at: Option<i64>,    // max(note.created_at) in ms epoch\n    pub last_seen_at: i64,\n    pub resolvable: bool,             // any note is resolvable\n    pub resolved: bool,               // all resolvable notes are resolved\n}\n\npub struct NormalizedNote {\n    pub gitlab_id: i64,\n    pub project_id: i64,\n    pub note_type: Option<String>,    // \"DiscussionNote\" | \"DiffNote\" | null\n    pub is_system: bool,              // from note.system\n    pub author_username: String,\n    pub body: String,\n    pub created_at: i64,              // ms epoch\n    pub updated_at: i64,              // ms epoch\n    pub last_seen_at: i64,\n    pub position: i32,                // 0-indexed array position\n    pub resolvable: bool,\n    pub resolved: bool,\n    pub resolved_by: Option<String>,\n    pub resolved_at: Option<i64>,\n}\n```\n\n### Functions\n\n```rust\npub fn transform_discussion(\n    gitlab_discussion: &GitLabDiscussion,\n    local_project_id: i64,\n    local_issue_id: i64,\n) -> NormalizedDiscussion\n\npub fn transform_notes(\n    gitlab_discussion: &GitLabDiscussion,\n    local_project_id: i64,\n) -> Vec<NormalizedNote>\n```\n\n## Acceptance Criteria\n\n- [ ] `NormalizedDiscussion` struct with all fields\n- [ ] `NormalizedNote` struct with all fields\n- [ ] `transform_discussion` computes first_note_at/last_note_at from notes array\n- [ ] `transform_discussion` computes resolvable (any note is resolvable)\n- [ ] `transform_discussion` computes resolved (all resolvable notes resolved)\n- [ ] `transform_notes` preserves array order via position field (0-indexed)\n- [ ] `transform_notes` maps system flag to is_system\n- [ ] Unit tests cover all computed fields\n\n## Files\n\n- src/gitlab/transformers/mod.rs (add `pub mod discussion;`)\n- src/gitlab/transformers/discussion.rs (create)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/discussion_transformer_tests.rs\n#[test] fn transforms_discussion_payload_to_normalized_schema()\n#[test] fn extracts_notes_array_from_discussion()\n#[test] fn sets_individual_note_flag_correctly()\n#[test] fn flags_system_notes_with_is_system_true()\n#[test] fn preserves_note_order_via_position_field()\n#[test] fn computes_first_note_at_and_last_note_at_correctly()\n#[test] fn computes_resolvable_and_resolved_status()\n```\n\nGREEN: Implement transform_discussion and transform_notes\n\nVERIFY: `cargo test discussion_transformer`\n\n## Edge Cases\n\n- Discussion with single note - first_note_at == last_note_at\n- All notes are system notes - still compute timestamps\n- No notes resolvable - resolvable=false, resolved=false\n- Mix of resolved/unresolved notes - resolved=false until all done","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.196079Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:27:11.485112Z","closed_at":"2026-01-25T22:27:11.485058Z","close_reason":"Implemented NormalizedDiscussion, NormalizedNote, transform_discussion, transform_notes with 9 passing unit tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1qf","depends_on_id":"bd-1np","type":"blocks","created_at":"2026-01-25T17:04:05.347218Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1qz","title":"[CP1] Database migration 002_issues.sql","description":"Create migration file with tables for issues, labels, issue_labels, discussions, and notes.\n\n## Tables\n\n### issues\n- id INTEGER PRIMARY KEY\n- gitlab_id INTEGER UNIQUE NOT NULL\n- project_id INTEGER NOT NULL REFERENCES projects(id)\n- iid INTEGER NOT NULL\n- title TEXT, description TEXT, state TEXT\n- author_username TEXT\n- created_at, updated_at, last_seen_at INTEGER (ms epoch UTC)\n- discussions_synced_for_updated_at INTEGER (watermark for dependent sync)\n- web_url TEXT\n- raw_payload_id INTEGER REFERENCES raw_payloads(id)\n\n### labels (name-only for CP1)\n- id INTEGER PRIMARY KEY\n- gitlab_id INTEGER (optional, for future Labels API)\n- project_id INTEGER NOT NULL REFERENCES projects(id)\n- name TEXT NOT NULL\n- color TEXT, description TEXT (nullable, deferred)\n- UNIQUE(project_id, name)\n\n### issue_labels (junction)\n- issue_id, label_id with CASCADE DELETE\n- Clear existing links before INSERT to handle removed labels\n\n### discussions\n- gitlab_discussion_id TEXT (string ID from API)\n- project_id, issue_id/merge_request_id FKs\n- noteable_type TEXT ('Issue' | 'MergeRequest')\n- individual_note INTEGER, first_note_at, last_note_at, last_seen_at\n- resolvable, resolved flags\n- CHECK constraint for Issue vs MR exclusivity\n\n### notes\n- gitlab_id INTEGER UNIQUE NOT NULL\n- discussion_id, project_id FKs\n- note_type, is_system, author_username, body\n- timestamps, position (array order)\n- resolution fields, DiffNote position fields\n\n## Indexes\n- idx_issues_project_updated, idx_issues_author, idx_issues_discussions_sync\n- uq_issues_project_iid, uq_labels_project_name\n- idx_issue_labels_label\n- uq_discussions_project_discussion_id, idx_discussions_issue/mr/last_note\n- idx_notes_discussion/author/system\n\nFiles: migrations/002_issues.sql\nDone when: Migration applies cleanly on top of 001_initial.sql, schema_version = 2","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:42:31.464544Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.685262Z","deleted_at":"2026-01-25T17:02:01.685258Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-1re","title":"[CP1] gi show issue command","description":"Show issue details with discussions.\n\nFlags:\n- --project=PATH (required if iid is ambiguous across projects)\n\nOutput:\n- Title, project, state, author, dates, labels, URL\n- Description text\n- All discussions with notes (formatted thread view)\n\nHandle ambiguity: If multiple projects have same iid, prompt for --project or show error.\n\nFiles: src/cli/commands/show.ts\nDone when: Issue detail view displays all fields including threaded discussions","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T15:20:29.826786Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.153211Z","deleted_at":"2026-01-25T15:21:35.153208Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-1s1","title":"[CP1] Integration tests for issue ingestion","description":"Full integration tests for issue ingestion module.\n\n## Tests (tests/issue_ingestion_tests.rs)\n\n- inserts_issues_into_database\n- creates_labels_from_issue_payloads\n- links_issues_to_labels_via_junction_table\n- removes_stale_label_links_on_resync\n- stores_raw_payload_for_each_issue\n- stores_raw_payload_for_each_discussion\n- updates_cursor_incrementally_per_page\n- resumes_from_cursor_on_subsequent_runs\n- handles_issues_with_no_labels\n- upserts_existing_issues_on_refetch\n- skips_discussion_refetch_for_unchanged_issues\n\n## Test Setup\n- tempfile::TempDir for isolated database\n- wiremock::MockServer for GitLab API\n- Mock handlers returning fixture data\n\nFiles: tests/issue_ingestion_tests.rs\nDone when: All integration tests pass with mocked GitLab","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:59:12.158586Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:02.109109Z","deleted_at":"2026-01-25T17:02:02.109105Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-1se","title":"Epic: Gate 2 - Cross-Reference Extraction","description":"Build entity relationship graph from structured APIs (closes_issues endpoint, state event source_merge_request_id) and best-effort system note parsing. The entity_references table is included in migration 011 (same migration as Gate 1 event tables).\n\nChildren: bd-czk (entity_references schema, folded into bd-hu3's migration), bd-8t4 (state event extraction), bd-3ia (closes_issues API), bd-1ji (system note parsing)","status":"open","priority":1,"issue_type":"feature","created_at":"2026-02-02T21:31:00.981132Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:44.441799Z","compaction_level":0,"original_size":0,"labels":["epic","gate-2","phase-b"],"dependencies":[{"issue_id":"bd-1se","depends_on_id":"bd-2zl","type":"blocks","created_at":"2026-02-02T21:32:43.028033Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1t4","title":"Epic: CP2 Gate C - Dependent Discussion Sync","description":"## Background\nGate C validates the dependent discussion sync with DiffNote position capture. This is critical for code review context preservation - without DiffNote positions, we lose the file/line context for review comments.\n\n## Acceptance Criteria (Pass/Fail)\n- [ ] Discussions fetched for MRs with updated_at > discussions_synced_for_updated_at\n- [ ] `SELECT COUNT(*) FROM discussions WHERE merge_request_id IS NOT NULL` > 0\n- [ ] DiffNotes have `position_new_path` populated (file path)\n- [ ] DiffNotes have `position_new_line` populated (line number)\n- [ ] DiffNotes have `position_type` populated (text/image/file)\n- [ ] DiffNotes have SHA triplet: `position_base_sha`, `position_start_sha`, `position_head_sha`\n- [ ] Multi-line DiffNotes have `position_line_range_start` and `position_line_range_end`\n- [ ] Unchanged MRs skip discussion refetch (watermark comparison works)\n- [ ] Watermark NOT advanced on HTTP error mid-pagination\n- [ ] Watermark NOT advanced on note timestamp parse failure\n- [ ] `gi show mr <iid>` displays DiffNote with file context `[path:line]`\n\n## Validation Script\n```bash\n#!/bin/bash\nset -e\n\nDB_PATH=\"${XDG_DATA_HOME:-$HOME/.local/share}/gitlab-inbox/db.sqlite3\"\n\necho \"=== Gate C: Dependent Discussion Sync ===\"\n\n# 1. Check discussion count for MRs\necho \"Step 1: Check MR discussion count...\"\nMR_DISC_COUNT=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM discussions WHERE merge_request_id IS NOT NULL;\")\necho \"  MR discussions: $MR_DISC_COUNT\"\n[ \"$MR_DISC_COUNT\" -gt 0 ] || { echo \"FAIL: No MR discussions found\"; exit 1; }\n\n# 2. Check note count\necho \"Step 2: Check note count...\"\nNOTE_COUNT=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM notes n\n  JOIN discussions d ON d.id = n.discussion_id\n  WHERE d.merge_request_id IS NOT NULL;\n\")\necho \"  MR notes: $NOTE_COUNT\"\n\n# 3. Check DiffNote position data\necho \"Step 3: Check DiffNote positions...\"\nDIFFNOTE_COUNT=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM notes WHERE position_new_path IS NOT NULL;\")\necho \"  DiffNotes with position: $DIFFNOTE_COUNT\"\n\n# 4. Sample DiffNote data\necho \"Step 4: Sample DiffNote data...\"\nsqlite3 \"$DB_PATH\" \"\n  SELECT \n    n.gitlab_id,\n    n.position_new_path,\n    n.position_new_line,\n    n.position_type,\n    SUBSTR(n.position_head_sha, 1, 7) as head_sha\n  FROM notes n\n  WHERE n.position_new_path IS NOT NULL\n  LIMIT 5;\n\"\n\n# 5. Check multi-line DiffNotes\necho \"Step 5: Check multi-line DiffNotes...\"\nMULTILINE_COUNT=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM notes \n  WHERE position_line_range_start IS NOT NULL \n    AND position_line_range_end IS NOT NULL\n    AND position_line_range_start != position_line_range_end;\n\")\necho \"  Multi-line DiffNotes: $MULTILINE_COUNT\"\n\n# 6. Check watermarks set\necho \"Step 6: Check watermarks...\"\nWATERMARKED=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM merge_requests \n  WHERE discussions_synced_for_updated_at IS NOT NULL;\n\")\necho \"  MRs with watermark set: $WATERMARKED\"\n\n# 7. Check last_seen_at for sweep pattern\necho \"Step 7: Check last_seen_at (sweep pattern)...\"\nsqlite3 \"$DB_PATH\" \"\n  SELECT \n    MIN(last_seen_at) as oldest,\n    MAX(last_seen_at) as newest\n  FROM discussions \n  WHERE merge_request_id IS NOT NULL;\n\"\n\n# 8. Test show command with DiffNote\necho \"Step 8: Find MR with DiffNotes for show test...\"\nMR_IID=$(sqlite3 \"$DB_PATH\" \"\n  SELECT DISTINCT m.iid\n  FROM merge_requests m\n  JOIN discussions d ON d.merge_request_id = m.id\n  JOIN notes n ON n.discussion_id = d.id\n  WHERE n.position_new_path IS NOT NULL\n  LIMIT 1;\n\")\nif [ -n \"$MR_IID\" ]; then\n  echo \"  Testing: gi show mr $MR_IID\"\n  gi show mr \"$MR_IID\" | head -50\nfi\n\n# 9. Re-run and verify skip count\necho \"Step 9: Re-run ingest (should skip unchanged MRs)...\"\ngi ingest --type=merge_requests\n# Should report \"Skipped discussion sync for N unchanged MRs\"\n\necho \"\"\necho \"=== Gate C: PASSED ===\"\n```\n\n## Atomicity Test (Manual - Kill Test)\n```bash\n# This tests that partial failure preserves data\n\n# 1. Get an MR with discussions\nMR_ID=$(sqlite3 \"$DB_PATH\" \"\n  SELECT m.id FROM merge_requests m\n  JOIN discussions d ON d.merge_request_id = m.id\n  LIMIT 1;\n\")\n\n# 2. Note current note count\nBEFORE=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM notes n\n  JOIN discussions d ON d.id = n.discussion_id\n  WHERE d.merge_request_id = $MR_ID;\n\")\necho \"Notes before: $BEFORE\"\n\n# 3. Note watermark\nWATERMARK_BEFORE=$(sqlite3 \"$DB_PATH\" \"\n  SELECT discussions_synced_for_updated_at FROM merge_requests WHERE id = $MR_ID;\n\")\necho \"Watermark before: $WATERMARK_BEFORE\"\n\n# 4. Force full sync and kill mid-run\ngi ingest --type=merge_requests --full &\nPID=$!\nsleep 3 && kill -9 $PID 2>/dev/null || true\nwait $PID 2>/dev/null || true\n\n# 5. Verify notes preserved (should be same or more, never less)\nAFTER=$(sqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM notes n\n  JOIN discussions d ON d.id = n.discussion_id\n  WHERE d.merge_request_id = $MR_ID;\n\")\necho \"Notes after kill: $AFTER\"\n[ \"$AFTER\" -ge \"$BEFORE\" ] || echo \"WARNING: Notes decreased - atomicity may be broken\"\n\n# 6. Note watermark should NOT have advanced if killed mid-pagination\nWATERMARK_AFTER=$(sqlite3 \"$DB_PATH\" \"\n  SELECT discussions_synced_for_updated_at FROM merge_requests WHERE id = $MR_ID;\n\")\necho \"Watermark after: $WATERMARK_AFTER\"\n```\n\n## Test Commands (Quick Verification)\n```bash\n# Check DiffNote data:\nsqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"\n  SELECT \n    (SELECT COUNT(*) FROM discussions WHERE merge_request_id IS NOT NULL) as mr_discussions,\n    (SELECT COUNT(*) FROM notes WHERE position_new_path IS NOT NULL) as diffnotes,\n    (SELECT COUNT(*) FROM merge_requests WHERE discussions_synced_for_updated_at IS NOT NULL) as watermarked;\n\"\n\n# Find MR with DiffNotes and show it:\ngi show mr $(sqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"\n  SELECT DISTINCT m.iid FROM merge_requests m\n  JOIN discussions d ON d.merge_request_id = m.id\n  JOIN notes n ON n.discussion_id = d.id\n  WHERE n.position_new_path IS NOT NULL LIMIT 1;\n\")\n```\n\n## Dependencies\nThis gate requires:\n- bd-3j6 (Discussion transformer with DiffNote position extraction)\n- bd-20h (MR discussion ingestion with atomicity guarantees)\n- bd-iba (Client pagination for MR discussions)\n- Gates A and B must pass first\n\n## Edge Cases\n- MRs without discussions: should sync successfully, just with 0 discussions\n- Discussions without DiffNotes: regular comments have NULL position fields\n- Deleted discussions in GitLab: sweep pattern should remove them locally\n- Invalid note timestamps: should NOT advance watermark, should log warning","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-26T22:06:01.769694Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:48:21.060017Z","closed_at":"2026-01-27T00:48:21.059974Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1t4","depends_on_id":"bd-20h","type":"blocks","created_at":"2026-01-26T22:08:55.778989Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1ta","title":"[CP1] Integration tests for pagination","description":"Integration tests for GitLab pagination with wiremock.\n\n## Tests (tests/pagination_tests.rs)\n\n### Page Navigation\n- fetches_all_pages_when_multiple_exist\n- respects_per_page_parameter\n- follows_x_next_page_header_until_empty\n- falls_back_to_empty_page_stop_if_headers_missing\n\n### Cursor Behavior\n- applies_cursor_rewind_for_tuple_semantics\n- clamps_negative_rewind_to_zero\n\n## Test Setup\n- Use wiremock::MockServer\n- Set up handlers for /api/v4/projects/:id/issues\n- Return x-next-page headers\n- Verify request params (updated_after, per_page)\n\nFiles: tests/pagination_tests.rs\nDone when: All pagination tests pass with mocked server","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:59:07.806593Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:02.038945Z","deleted_at":"2026-01-25T17:02:02.038939Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-1u1","title":"Implement document regenerator","description":"## Background\nThe document regenerator drains the dirty_sources queue, regenerating documents for each entry. It uses per-item transactions for crash safety, a triple-hash fast path to skip unchanged documents entirely (no writes at all), and a bounded batch loop that drains completely. Error recording includes backoff computation.\n\n## Approach\nCreate `src/documents/regenerator.rs` per PRD Section 6.3.\n\n**Core function:**\n```rust\npub fn regenerate_dirty_documents(conn: &Connection) -> Result<RegenerateResult>\n```\n\n**RegenerateResult:** { regenerated, unchanged, errored }\n\n**Algorithm (per PRD):**\n1. Loop: get_dirty_sources(conn) -> Vec<(SourceType, i64)>\n2. If empty, break (queue fully drained)\n3. For each (source_type, source_id):\n   a. Begin transaction\n   b. Call regenerate_one_tx(&tx, source_type, source_id) -> Result<bool>\n   c. If Ok(changed): clear_dirty_tx, commit, count regenerated or unchanged\n   d. If Err: record_dirty_error_tx (with backoff), commit, count errored\n\n**regenerate_one_tx (per PRD):**\n1. Extract document via extract_{type}_document(conn, source_id)\n2. If None (deleted): delete_document, return Ok(true)\n3. If Some(doc): call get_existing_hash() to check current state\n4. **If ALL THREE hashes match: return Ok(false) — skip ALL writes** (fast path)\n5. Otherwise: upsert_document with conditional label/path relinking\n6. Return Ok(content changed)\n\n**Helper functions (PRD-exact):**\n\n`get_existing_hash` — uses `optional()` to distinguish missing rows from DB errors:\n```rust\nfn get_existing_hash(\n    conn: &Connection,\n    source_type: SourceType,\n    source_id: i64,\n) -> Result<Option<String>> {\n    use rusqlite::OptionalExtension;\n    let hash: Option<String> = stmt\n        .query_row(params, |row| row.get(0))\n        .optional()?;  // IMPORTANT: Not .ok() — .ok() would hide real DB errors\n    Ok(hash)\n}\n```\n\n`get_document_id` — resolve document ID after upsert:\n```rust\nfn get_document_id(conn: &Connection, source_type: SourceType, source_id: i64) -> Result<i64>\n```\n\n`upsert_document` — checks existing triple hash before writing:\n```rust\nfn upsert_document(conn: &Connection, doc: &DocumentData) -> Result<()> {\n    // 1. Query existing (id, content_hash, labels_hash, paths_hash) via OptionalExtension\n    // 2. Triple-hash fast path: all match -> return Ok(())\n    // 3. Upsert document row (ON CONFLICT DO UPDATE)\n    // 4. Get doc_id (from existing or query after insert)\n    // 5. Only delete+reinsert labels if labels_hash changed\n    // 6. Only delete+reinsert paths if paths_hash changed\n}\n```\n\n**Key PRD detail — triple-hash fast path:**\n```rust\nif old_content_hash == &doc.content_hash\n    && old_labels_hash == &doc.labels_hash\n    && old_paths_hash == &doc.paths_hash\n{ return Ok(()); }  // Skip ALL writes — prevents WAL churn\n```\n\n**Error recording with backoff:**\nrecord_dirty_error_tx reads current attempt_count from DB, computes next_attempt_at via shared backoff utility:\n```rust\nlet next_attempt_at = crate::core::backoff::compute_next_attempt_at(now, attempt_count + 1);\n```\n\n**All internal functions use _tx suffix** (take &Transaction) for atomicity.\n\n## Acceptance Criteria\n- [ ] Queue fully drained (bounded batch loop until empty)\n- [ ] Per-item transactions (crash loses at most 1 doc)\n- [ ] Triple-hash fast path: ALL THREE hashes match -> skip ALL writes (return Ok(false))\n- [ ] Content change: upsert document, update labels/paths\n- [ ] Labels-only change: relabels but skips path writes (paths_hash unchanged)\n- [ ] Deleted entity: delete document (cascade handles FTS/labels/paths/embeddings)\n- [ ] get_existing_hash uses `.optional()` (not `.ok()`) to preserve DB errors\n- [ ] get_document_id resolves document ID after upsert\n- [ ] Error recording: increment attempt_count, compute next_attempt_at via backoff\n- [ ] FTS triggers fire on insert/update/delete (verified by trigger, not regenerator)\n- [ ] RegenerateResult counts accurate (regenerated, unchanged, errored)\n- [ ] Errors do not abort batch (log, increment, continue)\n- [ ] `cargo test regenerator` passes\n\n## Files\n- `src/documents/regenerator.rs` — new file\n- `src/documents/mod.rs` — add `pub use regenerator::regenerate_dirty_documents;`\n\n## TDD Loop\nRED: Tests requiring DB:\n- `test_creates_new_document` — dirty source -> document created\n- `test_skips_unchanged_triple_hash` — all 3 hashes match -> unchanged count incremented, no DB writes\n- `test_updates_changed_content` — content_hash mismatch -> updated\n- `test_updates_changed_labels_only` — content same but labels_hash different -> updated\n- `test_updates_changed_paths_only` — content same but paths_hash different -> updated\n- `test_deletes_missing_source` — source deleted -> document deleted\n- `test_drains_queue` — queue empty after regeneration\n- `test_error_records_backoff` — error -> attempt_count incremented, next_attempt_at set\n- `test_get_existing_hash_not_found` — returns Ok(None) for missing document\nGREEN: Implement regenerate_dirty_documents + all helpers\nVERIFY: `cargo test regenerator`\n\n## Edge Cases\n- Empty queue: return immediately with all-zero counts\n- Extractor error for one item: record_dirty_error_tx, commit, continue\n- Triple-hash prevents WAL churn on incremental syncs (most entities unchanged)\n- Labels change but content does not: labels_hash mismatch triggers upsert with label relinking\n- get_existing_hash on missing document: returns Ok(None) via .optional() (not DB error)\n- get_existing_hash on corrupt DB: propagates real DB error (not masked by .ok())","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:25:55.178825Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:41:29.942386Z","closed_at":"2026-01-30T17:41:29.942324Z","close_reason":"Implemented document regenerator with triple-hash fast path, queue draining, fail-soft error handling + 5 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1u1","depends_on_id":"bd-1yz","type":"blocks","created_at":"2026-01-30T15:29:16.020686Z","created_by":"tayloreernisse"},{"issue_id":"bd-1u1","depends_on_id":"bd-247","type":"blocks","created_at":"2026-01-30T15:29:15.982772Z","created_by":"tayloreernisse"},{"issue_id":"bd-1u1","depends_on_id":"bd-2fp","type":"blocks","created_at":"2026-01-30T15:29:16.055043Z","created_by":"tayloreernisse"}]}
+{"id":"bd-1uc","title":"Implement DB upsert functions for resource events","description":"## Background\nNeed to store fetched resource events into the three event tables created by migration 011. The existing DB pattern uses rusqlite prepared statements with named parameters. Timestamps from GitLab are ISO 8601 strings that need conversion to ms epoch UTC (matching the existing time.rs parse_datetime_to_ms function).\n\n## Approach\nCreate src/core/events_db.rs (new module) with three upsert functions:\n\n```rust\nuse rusqlite::Connection;\nuse super::error::Result;\n\n/// Upsert state events for an entity.\n/// Uses INSERT OR REPLACE keyed on UNIQUE(gitlab_id, project_id).\npub fn upsert_state_events(\n    conn: &Connection,\n    project_id: i64,          // local DB project id\n    entity_type: &str,        // \"issue\" | \"merge_request\"\n    entity_local_id: i64,     // local DB id of the issue/MR\n    events: &[GitLabStateEvent],\n) -> Result<usize>\n\n/// Upsert label events for an entity.\npub fn upsert_label_events(\n    conn: &Connection,\n    project_id: i64,\n    entity_type: &str,\n    entity_local_id: i64,\n    events: &[GitLabLabelEvent],\n) -> Result<usize>\n\n/// Upsert milestone events for an entity.\npub fn upsert_milestone_events(\n    conn: &Connection,\n    project_id: i64,\n    entity_type: &str,\n    entity_local_id: i64,\n    events: &[GitLabMilestoneEvent],\n) -> Result<usize>\n```\n\nEach function:\n1. Prepares INSERT OR REPLACE statement\n2. For each event, maps GitLab types to DB columns:\n   - `actor_gitlab_id` = event.user.map(|u| u.id)\n   - `actor_username` = event.user.map(|u| u.username.clone())\n   - `created_at` = parse_datetime_to_ms(&event.created_at)?\n   - Set issue_id or merge_request_id based on entity_type\n3. Returns count of upserted rows\n4. Wraps in a savepoint for atomicity per entity\n\nRegister module in src/core/mod.rs:\n```rust\npub mod events_db;\n```\n\n## Acceptance Criteria\n- [ ] All three upsert functions compile and handle all event fields\n- [ ] Upserts are idempotent (re-inserting same event doesn't duplicate)\n- [ ] Timestamps converted to ms epoch UTC via parse_datetime_to_ms\n- [ ] actor_gitlab_id and actor_username populated from event.user (handles None)\n- [ ] entity_type correctly maps to issue_id/merge_request_id (other is NULL)\n- [ ] source_merge_request_id populated for state events (iid from source_merge_request)\n- [ ] source_commit populated for state events\n- [ ] label_name populated for label events\n- [ ] milestone_title and milestone_id populated for milestone events\n- [ ] Returns upserted count\n\n## Files\n- src/core/events_db.rs (new)\n- src/core/mod.rs (add `pub mod events_db;`)\n\n## TDD Loop\nRED: tests/events_db_tests.rs (new):\n- `test_upsert_state_events_basic` - insert 3 events, verify count and data\n- `test_upsert_state_events_idempotent` - insert same events twice, verify no duplicates\n- `test_upsert_label_events_with_actor` - verify actor fields populated\n- `test_upsert_milestone_events_null_user` - verify user: null doesn't crash\n- `test_upsert_state_events_entity_exclusivity` - verify only one of issue_id/merge_request_id set\n\nSetup: create_test_db() helper that applies migrations 001-011, inserts a test project + issue + MR.\n\nGREEN: Implement the three functions\n\nVERIFY: `cargo test events_db -- --nocapture`\n\n## Edge Cases\n- parse_datetime_to_ms must handle GitLab's format: \"2024-03-15T10:30:00.000Z\" and \"2024-03-15T10:30:00.000+00:00\"\n- INSERT OR REPLACE will fire CASCADE deletes if there are FK references to these rows — currently no other table references event rows, so this is safe\n- entity_type must be validated (\"issue\" or \"merge_request\") — panic or error on invalid\n- source_merge_request field contains an MR ref object, not an ID — extract .iid for DB column","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:57.242549Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:42:12.906068Z","compaction_level":0,"original_size":0,"labels":["db","gate-1","phase-b"],"dependencies":[{"issue_id":"bd-1uc","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:57.246078Z","created_by":"tayloreernisse"},{"issue_id":"bd-1uc","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T21:31:57.247258Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1ut","title":"[CP0] Final validation - tests, lint, typecheck","description":"## Background\n\nFinal validation ensures everything works together before marking CP0 complete. This is the integration gate - all unit tests, integration tests, lint, and type checking must pass. Manual smoke tests verify the full user experience.\n\nReference: docs/prd/checkpoint-0.md sections \"Definition of Done\", \"Manual Smoke Tests\"\n\n## Approach\n\n**Automated checks:**\n```bash\n# All tests pass\nnpm run test\n\n# TypeScript strict mode\nnpm run build  # or: npx tsc --noEmit\n\n# ESLint with no errors\nnpm run lint\n```\n\n**Manual smoke tests (from PRD table):**\n\n| Command | Expected | Pass Criteria |\n|---------|----------|---------------|\n| `gi --help` | Command list | Shows all commands |\n| `gi version` | Version number | Shows installed version |\n| `gi init` | Interactive prompts | Creates valid config |\n| `gi init` (config exists) | Confirmation prompt | Warns before overwriting |\n| `gi init --force` | No prompt | Overwrites without asking |\n| `gi auth-test` | `Authenticated as @username` | Shows GitLab username |\n| `GITLAB_TOKEN=invalid gi auth-test` | Error message | Non-zero exit, clear error |\n| `gi doctor` | Status table | All required checks pass |\n| `gi doctor --json` | JSON object | Valid JSON, `success: true` |\n| `gi backup` | Backup path | Creates timestamped backup |\n| `gi sync-status` | No runs message | Stub output works |\n\n**Definition of Done gate items:**\n- [ ] `gi init` writes config to XDG path and validates projects against GitLab\n- [ ] `gi auth-test` succeeds with real PAT\n- [ ] `gi doctor` reports DB ok + GitLab ok\n- [ ] DB migrations apply; WAL + FK enabled; busy_timeout + synchronous set\n- [ ] App lock mechanism works (concurrent runs blocked)\n- [ ] All unit tests pass\n- [ ] All integration tests pass (mocked)\n- [ ] ESLint passes with no errors\n- [ ] TypeScript compiles with strict mode\n\n## Acceptance Criteria\n\n- [ ] `npm run test` exits 0 (all tests pass)\n- [ ] `npm run build` exits 0 (TypeScript compiles)\n- [ ] `npm run lint` exits 0 (no ESLint errors)\n- [ ] All 11 manual smoke tests pass\n- [ ] All 9 Definition of Done gate items verified\n\n## Files\n\nNo new files created. This bead verifies existing work.\n\n## TDD Loop\n\nThis IS the final verification step:\n\n```bash\n# Automated\nnpm run test\nnpm run build\nnpm run lint\n\n# Manual (requires GITLAB_TOKEN set with valid token)\ngi --help\ngi version\ngi init  # go through setup\ngi auth-test\ngi doctor\ngi doctor --json | jq .success  # should output true\ngi backup\ngi sync-status\ngi reset --confirm\ngi init  # re-setup\n```\n\n## Edge Cases\n\n- Test coverage should be reasonable (aim for 80%+ on core modules)\n- Integration tests may flake on CI - check MSW setup\n- Manual tests require real GitLab token - document in README\n- ESLint may warn vs error - only errors block\n- TypeScript noImplicitAny catches missed types","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:52.078907Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:37:51.858558Z","closed_at":"2026-01-25T03:37:51.858474Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1ut","depends_on_id":"bd-1cb","type":"blocks","created_at":"2026-01-24T16:13:11.184261Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ut","depends_on_id":"bd-1gu","type":"blocks","created_at":"2026-01-24T16:13:11.168637Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ut","depends_on_id":"bd-1kh","type":"blocks","created_at":"2026-01-24T16:13:11.219042Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ut","depends_on_id":"bd-38e","type":"blocks","created_at":"2026-01-24T16:13:11.150286Z","created_by":"tayloreernisse"},{"issue_id":"bd-1ut","depends_on_id":"bd-3kj","type":"blocks","created_at":"2026-01-24T16:13:11.200998Z","created_by":"tayloreernisse"}]}
+{"id":"bd-1v8","title":"Update robot-docs manifest with Phase B commands","description":"## Background\nThe `lore robot-docs` command outputs a machine-readable JSON manifest for agent self-discovery. New Phase B commands (timeline, file-history, trace, count references) must be registered so agents can discover and use them.\n\n## Approach\nIn src/cli/commands/robot_docs.rs, add entries to the commands array for:\n\n1. `timeline` - entity decision timeline\n   - args: entity type + IID, optional --project\n   - flags: --since, --until, --type (filter event types)\n   - output: chronological event list with actor/action/details\n\n2. `file-history` - file decision history\n   - args: file path\n   - flags: --since, --project\n   - output: MR/issue events touching the file\n\n3. `trace` - code trace from file to decisions\n   - args: file path or pattern\n   - flags: --depth, --since, --project\n   - output: connected graph of files → MRs → issues → cross-refs\n\n4. `count references` - cross-reference graph stats\n   - existing count command extended with new entity type\n\nEach entry follows the existing manifest structure with command name, description, args, flags, example, and output schema.\n\n## Acceptance Criteria\n- [ ] `lore robot-docs` JSON includes timeline command with args/flags/example\n- [ ] `lore robot-docs` JSON includes file-history command\n- [ ] `lore robot-docs` JSON includes trace command\n- [ ] `lore robot-docs` JSON includes references in count entity list\n- [ ] All new entries follow existing manifest schema\n- [ ] `lore robot-docs | jq '.commands[] | select(.name==\"timeline\")'` returns valid entry\n\n## Files\n- src/cli/commands/robot_docs.rs (add 4 command entries)\n\n## TDD Loop\nRED: tests/robot_docs_test.rs:\n- `test_robot_docs_includes_timeline` - parse JSON, assert timeline command exists\n- `test_robot_docs_includes_file_history` - assert file-history command exists\n- `test_robot_docs_includes_trace` - assert trace command exists\n- `test_robot_docs_count_includes_references` - assert count entity list includes references\n\nGREEN: Add command entries to the manifest builder\nREFACTOR: Extract command entry builder if pattern is repetitive\n\nVERIFY: `cargo test robot_docs -- --nocapture`\n\n## Edge Cases\n- Manifest must remain valid JSON (no trailing commas, proper escaping)\n- Command examples should use realistic flag combinations\n- Flag descriptions must match actual CLI behavior","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T22:43:07.859092Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:43:40.760221Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1v8","depends_on_id":"bd-1ht","type":"parent-child","created_at":"2026-02-02T22:43:40.760196Z","created_by":"tayloreernisse"},{"issue_id":"bd-1v8","depends_on_id":"bd-2ez","type":"blocks","created_at":"2026-02-02T22:43:33.990140Z","created_by":"tayloreernisse"},{"issue_id":"bd-1v8","depends_on_id":"bd-2n4","type":"blocks","created_at":"2026-02-02T22:43:33.937157Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1x6","title":"Implement lore sync CLI command","description":"## Background\nThe sync command is the unified orchestrator for the full pipeline: ingest -> generate-docs -> embed. It replaces the need to run three separate commands. It acquires a lock, runs each stage sequentially, and reports combined results. Individual stages can be skipped via flags (--no-embed, --no-docs). The command is designed for cron/scheduled execution. Individual commands (`lore generate-docs`, `lore embed`) still exist for manual recovery and debugging.\n\n## Approach\nCreate `src/cli/commands/sync.rs` per PRD Section 6.4.\n\n**IMPORTANT: run_sync is async** (embed_documents and search_hybrid are async).\n\n**Key types (PRD-exact):**\n```rust\n#[derive(Debug, Serialize)]\npub struct SyncResult {\n    pub issues_updated: usize,\n    pub mrs_updated: usize,\n    pub discussions_fetched: usize,\n    pub documents_regenerated: usize,\n    pub documents_embedded: usize,\n}\n\n#[derive(Debug, Default)]\npub struct SyncOptions {\n    pub full: bool,       // Reset cursors, fetch everything\n    pub force: bool,      // Override stale lock\n    pub no_embed: bool,   // Skip embedding step\n    pub no_docs: bool,    // Skip document regeneration\n}\n```\n\n**Core function (async, PRD-exact):**\n```rust\npub async fn run_sync(config: &Config, options: SyncOptions) -> Result<SyncResult>\n```\n\n**Pipeline (sequential steps per PRD):**\n1. Acquire app lock with heartbeat (via existing `src/core/lock.rs`)\n2. Ingest delta: fetch issues + MRs via cursor-based sync (calls existing ingestion orchestrator)\n   - Each upserted entity marked dirty via `mark_dirty_tx(&tx)` inside ingestion transaction\n3. Process `pending_discussion_fetches` queue (bounded)\n   - Discussion sweep uses CTE to capture stale IDs, then cascading deletes\n4. Regenerate documents from `dirty_sources` queue (unless --no-docs)\n5. Embed documents with changed content_hash (unless --no-embed; skipped gracefully if Ollama unavailable)\n6. Release lock, record sync_run\n\n**NOTE (PRD):** Rolling backfill window removed — the existing cursor + watermark design handles old issues with resumed activity. GitLab updates `updated_at` when new comments are added, so the cursor naturally picks up old issues that receive new activity.\n\n**CLI args (PRD-exact):**\n```rust\n#[derive(Args)]\npub struct SyncArgs {\n    /// Reset cursors, fetch everything\n    #[arg(long)]\n    full: bool,\n    /// Override stale lock\n    #[arg(long)]\n    force: bool,\n    /// Skip embedding step\n    #[arg(long)]\n    no_embed: bool,\n    /// Skip document regeneration\n    #[arg(long)]\n    no_docs: bool,\n}\n```\n\n**Human output:**\n```\nSync complete:\n  Issues updated:             42\n  MRs updated:                18\n  Discussions fetched:        56\n  Documents regenerated:      38\n  Documents embedded:         38\n  Elapsed: 2m 15s\n```\n\n**JSON output:**\n```json\n{\"ok\": true, \"data\": {...}, \"meta\": {\"elapsed_ms\": 135000}}\n```\n\n## Acceptance Criteria\n- [ ] Function is `async fn run_sync`\n- [ ] Takes `SyncOptions` struct (not separate params)\n- [ ] Returns `SyncResult` with flat fields (not nested sub-structs)\n- [ ] Full pipeline orchestrated: ingest -> discussion queue -> docs -> embed\n- [ ] --full resets cursors (passes through to ingest)\n- [ ] --force overrides stale sync lock\n- [ ] --no-embed skips embedding stage (Ollama not needed)\n- [ ] --no-docs skips document regeneration stage\n- [ ] Discussion queue processing bounded per run\n- [ ] Dirty sources marked inside ingestion transactions (via mark_dirty_tx)\n- [ ] Progress reporting: stage names + elapsed time\n- [ ] Lock acquired with heartbeat at start, released at end (even on error)\n- [ ] Embedding skipped gracefully if Ollama unavailable (warning, not error)\n- [ ] JSON summary in robot mode\n- [ ] Human-readable summary with elapsed time\n- [ ] `cargo build` succeeds\n\n## Files\n- `src/cli/commands/sync.rs` — new file\n- `src/cli/commands/mod.rs` — add `pub mod sync;`\n- `src/cli/mod.rs` — add SyncArgs, wire up sync subcommand\n- `src/main.rs` — add sync command handler (async dispatch)\n\n## TDD Loop\nRED: Integration test requiring full pipeline\nGREEN: Implement run_sync orchestration (async)\nVERIFY: `cargo build && cargo test sync`\n\n## Edge Cases\n- Ollama unavailable + --no-embed not set: sync should NOT fail — embed stage logs warning, returns 0 embedded\n- Lock already held: error unless --force (and lock is stale)\n- No dirty sources after ingest: regeneration stage returns 0 (not error)\n- --full with large dataset: keyset pagination prevents OFFSET degradation","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:27:09.577782Z","created_by":"tayloreernisse","updated_at":"2026-01-30T18:05:34.676100Z","closed_at":"2026-01-30T18:05:34.676035Z","close_reason":"Sync CLI: async run_sync orchestrator with 4-stage pipeline (ingest issues, ingest MRs, generate-docs, embed), SyncOptions/SyncResult, --full/--force/--no-embed/--no-docs flags, graceful Ollama degradation, human+JSON output, clean build, all tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1x6","depends_on_id":"bd-1i2","type":"blocks","created_at":"2026-01-30T15:29:35.287132Z","created_by":"tayloreernisse"},{"issue_id":"bd-1x6","depends_on_id":"bd-1je","type":"blocks","created_at":"2026-01-30T15:29:35.250622Z","created_by":"tayloreernisse"},{"issue_id":"bd-1x6","depends_on_id":"bd-2sx","type":"blocks","created_at":"2026-01-30T15:29:35.179059Z","created_by":"tayloreernisse"},{"issue_id":"bd-1x6","depends_on_id":"bd-38q","type":"blocks","created_at":"2026-01-30T15:29:35.213566Z","created_by":"tayloreernisse"},{"issue_id":"bd-1x6","depends_on_id":"bd-3qs","type":"blocks","created_at":"2026-01-30T15:29:35.144296Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1y8","title":"Implement chunk ID encoding module","description":"## Background\nsqlite-vec uses a single integer rowid for embeddings. To store multiple chunks per document, we encode (document_id, chunk_index) into a single rowid using a multiplier. This module is shared between the embedding pipeline (encode on write) and vector search (decode on read). The encoding scheme supports up to 1000 chunks per document.\n\n## Approach\nCreate `src/embedding/chunk_ids.rs`:\n\n```rust\n/// Multiplier for encoding (document_id, chunk_index) into a single rowid.\n/// Supports up to 1000 chunks per document (32M chars at 32k/chunk).\npub const CHUNK_ROWID_MULTIPLIER: i64 = 1000;\n\n/// Encode (document_id, chunk_index) into a sqlite-vec rowid.\n///\n/// rowid = document_id * CHUNK_ROWID_MULTIPLIER + chunk_index\npub fn encode_rowid(document_id: i64, chunk_index: i64) -> i64 {\n    document_id * CHUNK_ROWID_MULTIPLIER + chunk_index\n}\n\n/// Decode a sqlite-vec rowid back into (document_id, chunk_index).\npub fn decode_rowid(rowid: i64) -> (i64, i64) {\n    let document_id = rowid / CHUNK_ROWID_MULTIPLIER;\n    let chunk_index = rowid % CHUNK_ROWID_MULTIPLIER;\n    (document_id, chunk_index)\n}\n```\n\nAlso create the parent module `src/embedding/mod.rs`:\n```rust\npub mod chunk_ids;\n// Later beads add: pub mod ollama; pub mod pipeline;\n```\n\nUpdate `src/lib.rs`: add `pub mod embedding;`\n\n## Acceptance Criteria\n- [ ] `encode_rowid(42, 0)` == 42000\n- [ ] `encode_rowid(42, 5)` == 42005\n- [ ] `decode_rowid(42005)` == (42, 5)\n- [ ] Roundtrip: decode(encode(doc_id, chunk_idx)) == (doc_id, chunk_idx) for all valid inputs\n- [ ] CHUNK_ROWID_MULTIPLIER is 1000\n- [ ] `cargo test chunk_ids` passes\n\n## Files\n- `src/embedding/chunk_ids.rs` — new file\n- `src/embedding/mod.rs` — new file (module root)\n- `src/lib.rs` — add `pub mod embedding;`\n\n## TDD Loop\nRED: Tests in `#[cfg(test)] mod tests`:\n- `test_encode_single_chunk` — encode(1, 0) == 1000\n- `test_encode_multi_chunk` — encode(1, 5) == 1005\n- `test_decode_roundtrip` — property test over range of doc_ids and chunk_indices\n- `test_decode_zero_chunk` — decode(42000) == (42, 0)\n- `test_multiplier_value` — assert CHUNK_ROWID_MULTIPLIER == 1000\nGREEN: Implement encode_rowid, decode_rowid\nVERIFY: `cargo test chunk_ids`\n\n## Edge Cases\n- chunk_index >= 1000: not expected (documents that large would be pathological), but no runtime panic — just incorrect decode. The embedding pipeline caps chunks well below this.\n- document_id = 0: valid (encode returns chunk_index directly)","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:26:34.060769Z","created_by":"tayloreernisse","updated_at":"2026-01-30T16:51:59.048910Z","closed_at":"2026-01-30T16:51:59.048843Z","close_reason":"Completed: chunk_ids module with encode_rowid/decode_rowid, CHUNK_ROWID_MULTIPLIER=1000, 6 tests pass","compaction_level":0,"original_size":0}
 {"id":"bd-1yu","title":"[CP1] GitLab types for issues, discussions, notes","description":"Add TypeScript interfaces for GitLab API responses.\n\nTypes to add to src/gitlab/types.ts:\n- GitLabIssue: id, iid, project_id, title, description, state, timestamps, author, labels[], labels_details?, web_url\n- GitLabDiscussion: id (string), individual_note, notes[]\n- GitLabNote: id, type, body, author, timestamps, system, resolvable, resolved, resolved_by, resolved_at, position?\n\nFiles: src/gitlab/types.ts\nDone when: Types compile and match GitLab API documentation","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:19:00.558718Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.153996Z","deleted_at":"2026-01-25T15:21:35.153993Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-1yx","title":"Implement rename chain resolution for file-history","description":"## Background\nFile renames need to be resolved as bounded chains to find all historical paths. When a user queries for src/auth/oauth.rs, the system should also find MRs that touched src/auth/handler.rs if it was renamed. This is a core algorithm for both file-history and trace commands.\n\n## Approach\nCreate src/core/file_history.rs:\n\n```rust\n/// Result of rename chain resolution.\npub struct RenameChain {\n    /// All paths the file has been known by, ordered chronologically.\n    pub paths: Vec<String>,\n    /// The original query path.\n    pub query_path: String,\n}\n\n/// Resolve the rename chain for a file path.\n/// Starts with the query path and follows renames in both directions\n/// up to max_hops (default 10).\npub fn resolve_rename_chain(\n    conn: &Connection,\n    query_path: &str,\n    max_hops: usize,       // default 10, configurable\n) -> Result<RenameChain>\n```\n\nAlgorithm:\n```\n1. path_set = HashSet::from([query_path])\n2. for hop in 0..max_hops:\n3.   new_paths = query:\n        SELECT old_path, new_path FROM mr_file_changes\n        WHERE change_type = 'renamed'\n          AND (new_path IN path_set OR old_path IN path_set)\n4.   for (old, new) in new_paths:\n5.     add the \"other side\" (not already in path_set) to path_set\n6.   if no new paths discovered: break\n7. return ordered path set (attempt chronological ordering by MR date)\n```\n\nThe ordering uses MR merge/creation dates to present paths chronologically:\n```sql\nSELECT DISTINCT mfc.old_path, mfc.new_path, COALESCE(mr.merged_at, mr.created_at) as date\nFROM mr_file_changes mfc\nJOIN merge_requests mr ON mr.id = mfc.merge_request_id\nWHERE mfc.change_type = 'renamed'\n  AND (mfc.new_path IN (...) OR mfc.old_path IN (...))\nORDER BY date;\n```\n\nRegister in src/core/mod.rs: `pub mod file_history;`\n\n## Acceptance Criteria\n- [ ] Single rename resolved: query \"b.rs\" finds \"a.rs\" (renamed a→b)\n- [ ] Multi-hop: a→b→c chain resolved from any entry point\n- [ ] Cycle detection: a→b→a doesn't infinite loop\n- [ ] 10-hop cap enforced\n- [ ] Paths returned in chronological order\n- [ ] --no-follow-renames returns single-element chain (just query path)\n\n## Files\n- src/core/file_history.rs (new)\n- src/core/mod.rs (add `pub mod file_history;`)\n\n## TDD Loop\nRED: tests/file_history_tests.rs:\n- `test_rename_chain_single_hop` - a.rs renamed to b.rs, query b.rs → [a.rs, b.rs]\n- `test_rename_chain_multi_hop` - a→b→c, query c → [a, b, c]\n- `test_rename_chain_reverse_direction` - query a.rs also discovers b.rs\n- `test_rename_chain_cycle_detection` - a→b, b→a in different MRs → terminates\n- `test_rename_chain_hop_cap` - 15-hop chain with cap 10 → truncated at 10\n- `test_rename_chain_no_renames` - file never renamed → single-element chain\n\nSetup: create_test_db with migrations 001-012, seed mr_file_changes with renamed entries.\n\nGREEN: Implement resolve_rename_chain\n\nVERIFY: `cargo test file_history -- --nocapture`\n\n## Edge Cases\n- Same file renamed back and forth (a→b in MR !1, b→a in MR !2): cycle detection handles this\n- File renamed in two different projects: the query is project-scoped, so only same-project renames matter\n- old_path in mr_file_changes is only populated for change_type='renamed' — the WHERE clause already filters this\n- Path matching is exact (case-sensitive on Linux, should be consistent)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:34:08.985345Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:49:00.894547Z","compaction_level":0,"original_size":0,"labels":["gate-4","phase-b","query"],"dependencies":[{"issue_id":"bd-1yx","depends_on_id":"bd-14q","type":"parent-child","created_at":"2026-02-02T21:34:08.986730Z","created_by":"tayloreernisse"},{"issue_id":"bd-1yx","depends_on_id":"bd-1oo","type":"blocks","created_at":"2026-02-02T21:34:16.698782Z","created_by":"tayloreernisse"}]}
 {"id":"bd-1yz","title":"Implement MR document extraction","description":"## Background\nMR documents are similar to issue documents but include source/target branch information in the header. The extractor queries merge_requests and mr_labels tables. Like issue extraction, it produces a DocumentData struct for the regeneration pipeline.\n\n## Approach\nImplement `extract_mr_document()` in `src/documents/extractor.rs`:\n\n```rust\n/// Extract a searchable document from a merge request.\n/// Returns None if the MR has been deleted from the DB.\npub fn extract_mr_document(conn: &Connection, mr_id: i64) -> Result<Option<DocumentData>>\n```\n\n**SQL queries (from PRD Section 2.2):**\n```sql\n-- Main entity\nSELECT m.id, m.iid, m.title, m.description, m.state, m.author_username,\n       m.source_branch, m.target_branch,\n       m.created_at, m.updated_at, m.web_url,\n       p.path_with_namespace, p.id AS project_id\nFROM merge_requests m\nJOIN projects p ON p.id = m.project_id\nWHERE m.id = ?\n\n-- Labels\nSELECT l.name FROM mr_labels ml\nJOIN labels l ON l.id = ml.label_id\nWHERE ml.merge_request_id = ?\nORDER BY l.name\n```\n\n**Document format:**\n```\n[[MergeRequest]] !456: Implement JWT authentication\nProject: group/project-one\nURL: https://gitlab.example.com/group/project-one/-/merge_requests/456\nLabels: [\"feature\", \"auth\"]\nState: opened\nAuthor: @johndoe\nSource: feature/jwt-auth -> main\n\n--- Description ---\n\nThis MR implements JWT-based authentication...\n```\n\n**Key difference from issues:** The `Source:` line with `source_branch -> target_branch`.\n\n## Acceptance Criteria\n- [ ] Deleted MR returns Ok(None)\n- [ ] MR document has `[[MergeRequest]]` prefix with `!` before iid (not `#`)\n- [ ] Source line shows `source_branch -> target_branch`\n- [ ] Labels sorted alphabetically in JSON array\n- [ ] content_hash computed from full content_text\n- [ ] labels_hash computed from sorted labels\n- [ ] paths is empty (MR-level docs don't have DiffNote paths; those are on discussion docs)\n- [ ] `cargo test extract_mr` passes\n\n## Files\n- `src/documents/extractor.rs` — implement `extract_mr_document()`\n\n## TDD Loop\nRED: Tests in `#[cfg(test)] mod tests`:\n- `test_mr_document_format` — verify header matches PRD template with Source line\n- `test_mr_not_found` — returns Ok(None)\n- `test_mr_no_description` — header only\n- `test_mr_branch_info` — Source line correct\nGREEN: Implement extract_mr_document with SQL queries\nVERIFY: `cargo test extract_mr`\n\n## Edge Cases\n- MR with NULL description: skip \"--- Description ---\" section\n- MR with NULL source_branch or target_branch: omit Source line (shouldn't happen in practice)\n- Draft MRs: state field captures this, no special handling needed","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:25:45.521703Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:30:04.308781Z","closed_at":"2026-01-30T17:30:04.308598Z","close_reason":"Implemented extract_mr_document() with Source line, PRD format, and 5 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-1yz","depends_on_id":"bd-36p","type":"blocks","created_at":"2026-01-30T15:29:15.749264Z","created_by":"tayloreernisse"},{"issue_id":"bd-1yz","depends_on_id":"bd-hrs","type":"blocks","created_at":"2026-01-30T15:29:15.814729Z","created_by":"tayloreernisse"}]}
 {"id":"bd-208","title":"[CP1] Issue ingestion module","description":"## Background\n\nThe issue ingestion module fetches and stores issues with cursor-based incremental sync. It is the primary data ingestion component, establishing the pattern reused for MR ingestion in CP2. The module handles tuple-cursor semantics, raw payload storage, label extraction, and tracking which issues need discussion sync.\n\n## Approach\n\n### Module: src/ingestion/issues.rs\n\n### Key Structs\n\n```rust\n#[derive(Debug, Default)]\npub struct IngestIssuesResult {\n    pub fetched: usize,\n    pub upserted: usize,\n    pub labels_created: usize,\n    pub issues_needing_discussion_sync: Vec<IssueForDiscussionSync>,\n}\n\n#[derive(Debug, Clone)]\npub struct IssueForDiscussionSync {\n    pub local_issue_id: i64,\n    pub iid: i64,\n    pub updated_at: i64,  // ms epoch\n}\n```\n\n### Main Function\n\n```rust\npub async fn ingest_issues(\n    conn: &Connection,\n    client: &GitLabClient,\n    config: &Config,\n    project_id: i64,           // Local DB project ID\n    gitlab_project_id: i64,\n) -> Result<IngestIssuesResult>\n```\n\n### Logic (Step by Step)\n\n1. **Get current cursor** from sync_cursors table:\n```sql\nSELECT updated_at_cursor, tie_breaker_id\nFROM sync_cursors\nWHERE project_id = ? AND resource_type = 'issues'\n```\n\n2. **Call pagination method** with cursor rewind:\n```rust\nlet issues_stream = client.paginate_issues(\n    gitlab_project_id,\n    cursor.updated_at_cursor,\n    config.sync.cursor_rewind_seconds,\n);\n```\n\n3. **Apply local filtering** for tuple cursor semantics:\n```rust\n// Skip if issue.updated_at < cursor_updated_at\n// Skip if issue.updated_at == cursor_updated_at AND issue.gitlab_id <= cursor_gitlab_id\nfn passes_cursor_filter(issue: &GitLabIssue, cursor: &SyncCursor) -> bool {\n    if issue.updated_at < cursor.updated_at_cursor {\n        return false;\n    }\n    if issue.updated_at == cursor.updated_at_cursor \n        && issue.gitlab_id <= cursor.tie_breaker_id {\n        return false;\n    }\n    true\n}\n```\n\n4. **For each issue passing filter**:\n```rust\n// Begin transaction (unchecked_transaction for rusqlite)\nlet tx = conn.unchecked_transaction()?;\n\n// Store raw payload (compressed based on config)\nlet payload_id = store_raw_payload(&tx, &issue_json, config.storage.compress_raw_payloads)?;\n\n// Transform and upsert issue\nlet issue_row = transform_issue(&issue)?;\nupsert_issue(&tx, &issue_row, project_id, payload_id)?;\nlet local_issue_id = get_local_issue_id(&tx, project_id, issue.iid)?;\n\n// Clear existing label links (stale removal!)\ntx.execute(\"DELETE FROM issue_labels WHERE issue_id = ?\", [local_issue_id])?;\n\n// Extract and upsert labels\nfor label_name in &issue_row.label_names {\n    let label_id = upsert_label(&tx, project_id, label_name)?;\n    link_issue_label(&tx, local_issue_id, label_id)?;\n}\n\ntx.commit()?;\n```\n\n5. **Incremental cursor update** every 100 issues:\n```rust\nif batch_count % 100 == 0 {\n    update_sync_cursor(conn, project_id, \"issues\", last_updated_at, last_gitlab_id)?;\n}\n```\n\n6. **Final cursor update** after all issues processed\n\n7. **Determine issues needing discussion sync**:\n```sql\nSELECT id, iid, updated_at\nFROM issues\nWHERE project_id = ?\n  AND updated_at > COALESCE(discussions_synced_for_updated_at, 0)\n```\n\n### Helper Functions\n\n```rust\nfn store_raw_payload(conn, json: &Value, compress: bool) -> Result<i64>\nfn upsert_issue(conn, issue: &IssueRow, project_id: i64, payload_id: i64) -> Result<()>\nfn get_local_issue_id(conn, project_id: i64, iid: i64) -> Result<i64>\nfn upsert_label(conn, project_id: i64, name: &str) -> Result<i64>\nfn link_issue_label(conn, issue_id: i64, label_id: i64) -> Result<()>\nfn update_sync_cursor(conn, project_id: i64, resource: &str, updated_at: i64, gitlab_id: i64) -> Result<()>\n```\n\n### Critical Invariant\n\nStale label links MUST be removed on resync. The \"DELETE then INSERT\" pattern ensures GitLab reality is reflected locally. If an issue had labels [A, B] and now has [A, C], the B link must be removed.\n\n## Acceptance Criteria\n\n- [ ] `ingest_issues` returns IngestIssuesResult with all counts\n- [ ] Cursor fetched from sync_cursors at start\n- [ ] Cursor rewind applied before API call\n- [ ] Local filtering skips already-processed issues\n- [ ] Each issue wrapped in transaction for atomicity\n- [ ] Raw payload stored with correct compression\n- [ ] Issue upserted (INSERT OR REPLACE pattern)\n- [ ] Existing label links deleted before new links inserted\n- [ ] Labels upserted (INSERT OR IGNORE by project+name)\n- [ ] Cursor updated every 100 issues (crash recovery)\n- [ ] Final cursor update after all issues\n- [ ] issues_needing_discussion_sync populated correctly\n\n## Files\n\n- src/ingestion/mod.rs (add `pub mod issues;`)\n- src/ingestion/issues.rs (create)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/issue_ingestion_tests.rs\n#[tokio::test] async fn ingests_issues_from_stream()\n#[tokio::test] async fn applies_cursor_filter_correctly()\n#[tokio::test] async fn updates_cursor_every_100_issues()\n#[tokio::test] async fn stores_raw_payload_for_each_issue()\n#[tokio::test] async fn upserts_issues_correctly()\n\n// tests/label_linkage_tests.rs\n#[tokio::test] async fn extracts_and_stores_labels()\n#[tokio::test] async fn removes_stale_label_links_on_resync()\n#[tokio::test] async fn handles_empty_labels_array()\n\n// tests/discussion_eligibility_tests.rs\n#[tokio::test] async fn identifies_issues_needing_discussion_sync()\n#[tokio::test] async fn skips_issues_with_current_watermark()\n```\n\nGREEN: Implement ingest_issues with all helper functions\n\nVERIFY: `cargo test issue_ingestion && cargo test label_linkage && cargo test discussion_eligibility`\n\n## Edge Cases\n\n- Empty issues stream - return result with all zeros\n- Cursor at epoch 0 - fetch all issues (no filtering)\n- Issue with no labels - empty Vec, no label links created\n- Issue with 50+ labels - all should be linked\n- Crash mid-batch - cursor at last 100-boundary, some issues re-fetched\n- Label already exists - upsert via INSERT OR IGNORE\n- Same issue fetched twice (due to rewind) - upsert handles it","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.245404Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:52:38.003964Z","closed_at":"2026-01-25T22:52:38.003868Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-208","depends_on_id":"bd-2iq","type":"blocks","created_at":"2026-01-25T17:04:05.425224Z","created_by":"tayloreernisse"},{"issue_id":"bd-208","depends_on_id":"bd-3nd","type":"blocks","created_at":"2026-01-25T17:04:05.450341Z","created_by":"tayloreernisse"},{"issue_id":"bd-208","depends_on_id":"bd-xhz","type":"blocks","created_at":"2026-01-25T17:04:05.473203Z","created_by":"tayloreernisse"}]}
+{"id":"bd-20e","title":"Define TimelineEvent model and TimelineEventType enum","description":"## Background\nThe timeline needs a unified event model that spans multiple source tables (resource events, issues, MRs, notes). This is a read-time virtual event stream — not a separate stored table. The spec (§3.3) defines the exact shapes.\n\n## Approach\nCreate src/core/timeline.rs with the core types:\n\n```rust\nuse serde::Serialize;\n\n/// A single event in a timeline query result.\n#[derive(Debug, Clone, Serialize)]\npub struct TimelineEvent {\n    pub timestamp: i64,              // ms epoch UTC\n    pub entity_type: String,         // \"issue\" | \"merge_request\"\n    pub entity_iid: i64,\n    pub project_path: String,\n    pub event_type: TimelineEventType,\n    pub summary: String,             // human-readable one-liner\n    pub actor: Option<String>,       // username\n    pub url: Option<String>,\n    pub is_seed: bool,               // matched by keyword (vs. expanded via reference)\n}\n\n/// Types of events in the timeline.\n#[derive(Debug, Clone, Serialize)]\n#[serde(tag = \"type\", rename_all = \"snake_case\")]\npub enum TimelineEventType {\n    Created,\n    StateChanged { state: String },\n    LabelAdded { label: String },\n    LabelRemoved { label: String },\n    MilestoneSet { milestone: String },\n    MilestoneRemoved { milestone: String },\n    Merged,\n    NoteEvidence {\n        note_id: i64,\n        snippet: String,              // first ~200 chars\n        discussion_id: Option<i64>,\n    },\n    CrossReferenced { target: String },\n}\n\nimpl TimelineEvent {\n    /// Create a summary string for human output.\n    pub fn format_summary(&self) -> String { ... }\n}\n\nimpl Ord for TimelineEvent {\n    /// Sort chronologically by timestamp, then by entity_iid for stable tiebreak.\n    fn cmp(&self, other: &Self) -> Ordering {\n        self.timestamp.cmp(&other.timestamp)\n            .then_with(|| self.entity_iid.cmp(&other.entity_iid))\n    }\n}\nimpl PartialOrd for TimelineEvent { ... }\nimpl Eq for TimelineEvent {}\nimpl PartialEq for TimelineEvent { ... }\n\n/// Result of a timeline query.\npub struct TimelineResult {\n    pub query: String,\n    pub events: Vec<TimelineEvent>,\n    pub seed_entities: Vec<EntityRef>,\n    pub expanded_entities: Vec<ExpandedEntityRef>,\n    pub unresolved_references: Vec<UnresolvedRef>,\n}\n\n#[derive(Debug, Clone, Serialize)]\npub struct EntityRef {\n    pub entity_type: String,\n    pub iid: i64,\n    pub project: String,\n}\n\n#[derive(Debug, Clone, Serialize)]\npub struct ExpandedEntityRef {\n    pub entity_type: String,\n    pub iid: i64,\n    pub project: String,\n    pub depth: usize,\n    pub via: ExpansionProvenance,\n}\n\n#[derive(Debug, Clone, Serialize)]\npub struct ExpansionProvenance {\n    pub from: EntityRef,\n    pub reference_type: String,\n    pub source_method: String,\n}\n\n#[derive(Debug, Clone, Serialize)]\npub struct UnresolvedRef {\n    pub source: EntityRef,\n    pub target_project: String,\n    pub target_type: String,\n    pub target_iid: i64,\n    pub reference_type: String,\n}\n```\n\nRegister in src/core/mod.rs: `pub mod timeline;`\n\n## Acceptance Criteria\n- [ ] All types compile with Serialize derive\n- [ ] TimelineEventType serde serializes with tagged union (type field)\n- [ ] TimelineEvent Ord sorts chronologically with stable tiebreak\n- [ ] format_summary produces readable strings for each event type\n- [ ] All helper types (EntityRef, ExpandedEntityRef, UnresolvedRef) defined\n\n## Files\n- src/core/timeline.rs (new)\n- src/core/mod.rs (add `pub mod timeline;`)\n\n## TDD Loop\nRED: tests/timeline_types_tests.rs:\n- `test_timeline_event_sort_chronological` - events at different timestamps sort correctly\n- `test_timeline_event_sort_stable_tiebreak` - same timestamp, different iids, stable order\n- `test_timeline_event_type_serialize` - verify JSON tag serialization for each variant\n- `test_format_summary_created` - verify human string for Created\n- `test_format_summary_state_changed` - verify \"closed\" / \"reopened\" etc.\n- `test_format_summary_note_evidence` - verify snippet truncation\n\nGREEN: Define all types and implement traits\n\nVERIFY: `cargo test timeline_types -- --nocapture`\n\n## Edge Cases\n- TimelineEventType::NoteEvidence snippet should be truncated at ~200 chars with \"...\" suffix — handle multi-byte UTF-8 correctly (don't split in middle of char)\n- serde tag=\"type\" may conflict if any variant has a field named \"type\" — none do, so safe\n- Ord comparison on timestamps should handle equal timestamps gracefully — iid tiebreak ensures stability","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:08.569126Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:45:25.651058Z","compaction_level":0,"original_size":0,"labels":["gate-3","phase-b","types"],"dependencies":[{"issue_id":"bd-20e","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:08.573079Z","created_by":"tayloreernisse"}]}
 {"id":"bd-20h","title":"Implement MR discussion ingestion module","description":"## Background\nMR discussion ingestion with critical atomicity guarantees. Parse notes BEFORE destructive DB operations to prevent data loss. Watermark ONLY advanced on full success.\n\n## Approach\nCreate `src/ingestion/mr_discussions.rs` with:\n1. `IngestMrDiscussionsResult` - Per-MR stats\n2. `ingest_mr_discussions()` - Main function with atomicity guarantees\n3. Upsert + sweep pattern for notes (not delete-all-then-insert)\n4. Sync health telemetry for debugging failures\n\n## Files\n- `src/ingestion/mr_discussions.rs` - New module\n- `tests/mr_discussion_ingestion_tests.rs` - Integration tests\n\n## Acceptance Criteria\n- [ ] `IngestMrDiscussionsResult` has: discussions_fetched, discussions_upserted, notes_upserted, notes_skipped_bad_timestamp, diffnotes_count, pagination_succeeded\n- [ ] `ingest_mr_discussions()` returns `Result<IngestMrDiscussionsResult>`\n- [ ] CRITICAL: Notes parsed BEFORE any DELETE operations\n- [ ] CRITICAL: Watermark NOT advanced if `pagination_succeeded == false`\n- [ ] CRITICAL: Watermark NOT advanced if any note parse fails\n- [ ] Upsert + sweep pattern using `last_seen_at`\n- [ ] Stale discussions/notes removed only on full success\n- [ ] Selective raw payload storage (skip system notes without position)\n- [ ] Sync health telemetry recorded on failure\n- [ ] `does_not_advance_discussion_watermark_on_partial_failure` test passes\n- [ ] `atomic_note_replacement_preserves_data_on_parse_failure` test passes\n\n## TDD Loop\nRED: `cargo test does_not_advance_watermark` -> test fails\nGREEN: Add ingestion with atomicity guarantees\nVERIFY: `cargo test mr_discussion_ingestion`\n\n## Main Function\n```rust\npub async fn ingest_mr_discussions(\n    conn: &Connection,\n    client: &GitLabClient,\n    config: &Config,\n    project_id: i64,\n    gitlab_project_id: i64,\n    mr_iid: i64,\n    local_mr_id: i64,\n    mr_updated_at: i64,\n) -> Result<IngestMrDiscussionsResult>\n```\n\n## CRITICAL: Atomic Note Replacement\n```rust\n// Record sync start time for sweep\nlet run_seen_at = now_ms();\n\nwhile let Some(discussion_result) = stream.next().await {\n    let discussion = match discussion_result {\n        Ok(d) => d,\n        Err(e) => {\n            result.pagination_succeeded = false;\n            break; // Stop but don't advance watermark\n        }\n    };\n    \n    // CRITICAL: Parse BEFORE destructive operations\n    let notes = match transform_notes_with_diff_position(&discussion, project_id) {\n        Ok(notes) => notes,\n        Err(e) => {\n            warn!(\"Note transform failed; preserving existing notes\");\n            result.notes_skipped_bad_timestamp += discussion.notes.len();\n            result.pagination_succeeded = false;\n            continue; // Skip this discussion, don't delete existing\n        }\n    };\n    \n    // Only NOW start transaction (after parse succeeded)\n    let tx = conn.unchecked_transaction()?;\n    \n    // Upsert discussion with run_seen_at\n    // Upsert notes with run_seen_at (not delete-all)\n    \n    tx.commit()?;\n}\n```\n\n## Stale Data Sweep (only on success)\n```rust\nif result.pagination_succeeded {\n    // Sweep stale discussions\n    conn.execute(\n        \"DELETE FROM discussions\n         WHERE project_id = ? AND merge_request_id = ?\n           AND last_seen_at < ?\",\n        params![project_id, local_mr_id, run_seen_at],\n    )?;\n    \n    // Sweep stale notes\n    conn.execute(\n        \"DELETE FROM notes\n         WHERE discussion_id IN (\n           SELECT id FROM discussions\n           WHERE project_id = ? AND merge_request_id = ?\n         )\n         AND last_seen_at < ?\",\n        params![project_id, local_mr_id, run_seen_at],\n    )?;\n}\n```\n\n## Watermark Update (ONLY on success)\n```rust\nif result.pagination_succeeded {\n    mark_discussions_synced(conn, local_mr_id, mr_updated_at)?;\n    clear_sync_health_error(conn, local_mr_id)?;\n} else {\n    record_sync_health_error(conn, local_mr_id, \"Pagination incomplete or parse failure\")?;\n    warn!(\"Watermark NOT advanced; will retry on next sync\");\n}\n```\n\n## Selective Payload Storage\n```rust\n// Only store payload for DiffNotes and non-system notes\nlet should_store_note_payload =\n    !note.is_system() ||\n    note.position_new_path().is_some() ||\n    note.position_old_path().is_some();\n```\n\n## Integration Tests (CRITICAL)\n```rust\n#[tokio::test]\nasync fn does_not_advance_discussion_watermark_on_partial_failure() {\n    // Setup: MR with updated_at > discussions_synced_for_updated_at\n    // Mock: Page 1 returns OK, Page 2 returns 500\n    // Assert: discussions_synced_for_updated_at unchanged\n}\n\n#[tokio::test]\nasync fn does_not_advance_discussion_watermark_on_note_parse_failure() {\n    // Setup: Existing notes in DB\n    // Mock: Discussion with note having invalid created_at\n    // Assert: Original notes preserved, watermark unchanged\n}\n\n#[tokio::test]\nasync fn atomic_note_replacement_preserves_data_on_parse_failure() {\n    // Setup: Discussion with 3 valid notes\n    // Mock: Updated discussion where note 2 has bad timestamp\n    // Assert: All 3 original notes still in DB\n}\n```\n\n## Edge Cases\n- HTTP error mid-pagination: preserve existing data, log error, no watermark advance\n- Invalid note timestamp: skip discussion, preserve existing notes\n- System notes without position: don't store raw payload (saves space)\n- Empty discussion: still upsert discussion record, no notes","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:42.335714Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:22:43.207057Z","closed_at":"2026-01-27T00:22:43.206996Z","close_reason":"Implemented MR discussion ingestion module with full atomicity guarantees:\n- IngestMrDiscussionsResult with all required fields\n- parse-before-destructive pattern (transform notes before DB ops)\n- Upsert + sweep pattern with last_seen_at timestamps\n- Watermark advanced ONLY on full pagination success\n- Selective payload storage (skip system notes without position)\n- Sync health telemetry for failure debugging\n- All 163 tests passing","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-20h","depends_on_id":"bd-3ir","type":"blocks","created_at":"2026-01-26T22:08:54.649094Z","created_by":"tayloreernisse"},{"issue_id":"bd-20h","depends_on_id":"bd-3j6","type":"blocks","created_at":"2026-01-26T22:08:54.686066Z","created_by":"tayloreernisse"},{"issue_id":"bd-20h","depends_on_id":"bd-iba","type":"blocks","created_at":"2026-01-26T22:08:54.722746Z","created_by":"tayloreernisse"}]}
 {"id":"bd-221","title":"Create migration 008_fts5.sql","description":"## Background\nFTS5 (Full-Text Search 5) provides the lexical search backbone for Gate A. The virtual table + triggers keep the FTS index in sync with the documents table automatically. This migration must be applied AFTER migration 007 (documents table exists). The trigger design handles NULL titles via COALESCE and only rebuilds the FTS entry when searchable text actually changes (not metadata-only updates).\n\n## Approach\nCreate `migrations/008_fts5.sql` with the exact SQL from PRD Section 1.2:\n\n1. **Virtual table:** `documents_fts` using FTS5 with porter stemmer, prefix indexes (2,3,4), external content backed by `documents` table\n2. **Insert trigger:** `documents_ai` — inserts into FTS on document insert, uses COALESCE(title, '') for NULL safety\n3. **Delete trigger:** `documents_ad` — removes from FTS on document delete using the FTS5 delete command syntax\n4. **Update trigger:** `documents_au` — only fires when `title` or `content_text` changes (WHEN clause), performs delete-then-insert to update FTS\n\nRegister migration 8 in `src/core/db.rs` MIGRATIONS array.\n\n**Critical detail:** The COALESCE is required because FTS5 external-content tables require exact value matching for delete operations. If NULL was inserted, the delete trigger couldn't match it (NULL != NULL in SQL).\n\n## Acceptance Criteria\n- [ ] `migrations/008_fts5.sql` file exists\n- [ ] `documents_fts` virtual table created with `tokenize='porter unicode61'` and `prefix='2 3 4'`\n- [ ] `content='documents'` and `content_rowid='id'` set (external content mode)\n- [ ] Insert trigger `documents_ai` fires on document insert with COALESCE(title, '')\n- [ ] Delete trigger `documents_ad` fires on document delete using FTS5 delete command\n- [ ] Update trigger `documents_au` only fires when `old.title IS NOT new.title OR old.content_text != new.content_text`\n- [ ] Prefix search works: query `auth*` matches \"authentication\"\n- [ ] After bulk insert of N documents, `SELECT count(*) FROM documents_fts` returns N\n- [ ] Schema version 8 recorded in schema_version table\n- [ ] `cargo test migration_tests` passes\n\n## Files\n- `migrations/008_fts5.sql` — new file (copy exact SQL from PRD Section 1.2)\n- `src/core/db.rs` — add migration 8 to MIGRATIONS array\n\n## TDD Loop\nRED: Register migration in db.rs, `cargo test migration_tests` fails (SQL file missing)\nGREEN: Create `008_fts5.sql` with all triggers\nVERIFY: `cargo test migration_tests && cargo build`\n\n## Edge Cases\n- Metadata-only updates (e.g., changing `updated_at` or `labels_hash`) must NOT trigger FTS rebuild — the WHEN clause prevents this\n- NULL titles must use COALESCE to empty string in both insert and delete triggers\n- The update trigger does delete+insert (not FTS5 'delete' + regular insert atomically) — this is the correct FTS5 pattern for content changes","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:25:25.763146Z","created_by":"tayloreernisse","updated_at":"2026-01-30T16:56:13.131830Z","closed_at":"2026-01-30T16:56:13.131771Z","close_reason":"Completed: migration 008_fts5.sql with FTS5 virtual table, 3 sync triggers (insert/delete/update with COALESCE NULL safety), prefix search, registered in db.rs, cargo build + tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-221","depends_on_id":"bd-hrs","type":"blocks","created_at":"2026-01-30T15:29:15.574576Z","created_by":"tayloreernisse"}]}
 {"id":"bd-227","title":"[CP1] gi count issues/discussions/notes commands","description":"Count entities in the database.\n\n## Module\nsrc/cli/commands/count.rs\n\n## Clap Definition\nCount {\n    #[arg(value_parser = [\"issues\", \"mrs\", \"discussions\", \"notes\"])]\n    entity: String,\n    \n    #[arg(long, value_parser = [\"issue\", \"mr\"])]\n    r#type: Option<String>,\n}\n\n## Commands\n- gi count issues → 'Issues: N'\n- gi count discussions → 'Discussions: N'\n- gi count discussions --type=issue → 'Issue Discussions: N'\n- gi count notes → 'Notes: N (excluding M system)'\n- gi count notes --type=issue → 'Issue Notes: N (excluding M system)'\n\n## Implementation\n- Simple COUNT(*) queries\n- For notes, also count WHERE is_system = 1 for system note count\n- Filter by noteable_type when --type specified\n\nFiles: src/cli/commands/count.rs\nDone when: Counts match expected values from GitLab","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:58:25.648805Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.920135Z","deleted_at":"2026-01-25T17:02:01.920129Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-247","title":"Implement issue document extraction","description":"## Background\nIssue documents are the simplest document type — a structured header + description text. The extractor queries the existing issues and issue_labels tables (populated by ingestion) and assembles a DocumentData struct. This is one of three entity-specific extractors (issue, MR, discussion) that feed the document regeneration pipeline.\n\n## Approach\nImplement `extract_issue_document()` in `src/documents/extractor.rs`:\n\n```rust\n/// Extract a searchable document from an issue.\n/// Returns None if the issue has been deleted from the DB.\npub fn extract_issue_document(conn: &Connection, issue_id: i64) -> Result<Option<DocumentData>>\n```\n\n**SQL queries (from PRD Section 2.2):**\n```sql\n-- Main entity\nSELECT i.id, i.iid, i.title, i.description, i.state, i.author_username,\n       i.created_at, i.updated_at, i.web_url,\n       p.path_with_namespace, p.id AS project_id\nFROM issues i\nJOIN projects p ON p.id = i.project_id\nWHERE i.id = ?\n\n-- Labels\nSELECT l.name FROM issue_labels il\nJOIN labels l ON l.id = il.label_id\nWHERE il.issue_id = ?\nORDER BY l.name\n```\n\n**Document format:**\n```\n[[Issue]] #234: Authentication redesign\nProject: group/project-one\nURL: https://gitlab.example.com/group/project-one/-/issues/234\nLabels: [\"bug\", \"auth\"]\nState: opened\nAuthor: @johndoe\n\n--- Description ---\n\nWe need to modernize our authentication system...\n```\n\n**Implementation steps:**\n1. Query issue row — if not found, return Ok(None)\n2. Query labels via junction table\n3. Format header with [[Issue]] prefix\n4. Compute content_hash via compute_content_hash()\n5. Compute labels_hash via compute_list_hash()\n6. paths is always empty for issues (paths are only for DiffNote discussions)\n7. Return DocumentData with all fields populated\n\n## Acceptance Criteria\n- [ ] Deleted issue (not in DB) returns Ok(None)\n- [ ] Issue with no description: content_text has header only (no \"--- Description ---\" section)\n- [ ] Issue with no labels: Labels line shows \"[]\"\n- [ ] Issue with labels: Labels line shows sorted JSON array\n- [ ] content_hash is SHA-256 of the full content_text\n- [ ] labels_hash is SHA-256 of sorted label names joined by newline\n- [ ] paths_hash is empty string hash (issues have no paths)\n- [ ] project_id comes from the JOIN with projects table\n- [ ] `cargo test extract_issue` passes\n\n## Files\n- `src/documents/extractor.rs` — implement `extract_issue_document()`\n\n## TDD Loop\nRED: Test in `#[cfg(test)] mod tests`:\n- `test_issue_document_format` — verify header format matches PRD template\n- `test_issue_not_found` — returns Ok(None) for nonexistent issue_id\n- `test_issue_no_description` — no description section when description is NULL\n- `test_issue_labels_sorted` — labels appear in alphabetical order\n- `test_issue_hash_deterministic` — same issue produces same content_hash\nGREEN: Implement extract_issue_document with SQL queries\nVERIFY: `cargo test extract_issue`\n\n## Edge Cases\n- Issue with NULL description: skip \"--- Description ---\" section entirely\n- Issue with empty string description: include section but with empty body\n- Issue with very long description: no truncation here (hard cap applied by caller)\n- Labels with special characters (quotes, commas): JSON array handles escaping","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:25:45.490145Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:28:13.974948Z","closed_at":"2026-01-30T17:28:13.974891Z","close_reason":"Implemented extract_issue_document() with SQL queries, PRD-compliant format, and 7 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-247","depends_on_id":"bd-36p","type":"blocks","created_at":"2026-01-30T15:29:15.677223Z","created_by":"tayloreernisse"},{"issue_id":"bd-247","depends_on_id":"bd-hrs","type":"blocks","created_at":"2026-01-30T15:29:15.712739Z","created_by":"tayloreernisse"}]}
-{"id":"bd-25s","title":"robot-docs: Add Ollama dependency discovery to manifest","description":"An agent reading robot-docs has no way to discover that semantic search requires Ollama, what model to pull, or that lexical search works without it. The error messages in embed/search are actionable enough for self-correction, but proactive discovery would be better.\n\nProposed: add a 'dependencies' section to the robot-docs JSON output listing optional deps with install commands and setup instructions. Example:\n\n{\"ollama\": {\"required_for\": [\"embed\", \"search --mode semantic|hybrid\"], \"install\": \"brew install ollama || curl -fsSL https://ollama.com/install.sh | sh\", \"setup\": \"ollama serve && ollama pull nomic-embed-text\", \"optional\": true, \"note\": \"Lexical search works without Ollama\"}}\n\nExisting coverage (sufficient for now):\n- README.md mentions Ollama as a dependency\n- doctor command checks Ollama health at runtime\n- error.rs has actionable suggestions: 'ollama serve', 'ollama pull nomic-embed-text'\n- sync gracefully degrades when Ollama is unavailable\n\nGap: none of this surfaces in robot-docs, the agent self-discovery entry point.","status":"open","priority":4,"issue_type":"feature","created_at":"2026-01-30T20:26:43.169688Z","created_by":"tayloreernisse","updated_at":"2026-01-30T20:26:43.173276Z","compaction_level":0,"original_size":0,"labels":["enhancement","robot-mode"]}
+{"id":"bd-25s","title":"robot-docs: Add Ollama dependency discovery to manifest","description":"## Background\nAn agent reading robot-docs has no way to discover that semantic search requires Ollama, what model to pull, or that lexical search works without it. The error messages in embed/search are actionable enough for self-correction, but proactive discovery would be better.\n\n## Approach\nAdd a 'dependencies' section to the robot-docs JSON output in the command that generates it. The section lists optional external dependencies with install commands, setup instructions, and which commands require them.\n\nIn the robot-docs handler (likely src/cli/commands/robot_docs.rs or wherever the robot-docs JSON is built):\n\n```rust\n\"dependencies\": {\n    \"ollama\": {\n        \"required_for\": [\"embed\", \"search --mode semantic\", \"search --mode hybrid\"],\n        \"not_required_for\": [\"search --mode lexical\", \"list\", \"show\", \"count\", \"timeline\", \"file-history\", \"trace\"],\n        \"install\": {\n            \"macos\": \"brew install ollama\",\n            \"linux\": \"curl -fsSL https://ollama.com/install.sh | sh\",\n            \"check\": \"ollama --version\"\n        },\n        \"setup\": [\"ollama serve\", \"ollama pull nomic-embed-text\"],\n        \"optional\": true,\n        \"note\": \"Lexical search, temporal queries, and all non-embedding features work without Ollama. Only semantic/hybrid search and the embed command require it.\"\n    }\n}\n```\n\nLocate the existing robot-docs generation code and add this section to the top-level JSON object.\n\n## Acceptance Criteria\n- [ ] `lore robot-docs` JSON includes a \"dependencies\" section\n- [ ] Ollama listed with required_for commands, install commands, and setup steps\n- [ ] optional: true flag present\n- [ ] Note clarifying what works without Ollama\n- [ ] Existing robot-docs fields unchanged\n\n## Files\n- src/cli/commands/robot_docs.rs (or wherever robot-docs JSON is generated)\n\n## TDD Loop\nRED: tests/robot_docs_tests.rs:\n- `test_robot_docs_includes_dependencies` - parse JSON, verify \"dependencies\" key exists\n- `test_robot_docs_ollama_dependency` - verify ollama entry has required_for, install, setup\n\nGREEN: Add the dependencies section to robot-docs output\n\nVERIFY: `cargo test robot_docs -- --nocapture`\n\n## Edge Cases\n- Keep the dependencies section extensible (HashMap or Vec) for future deps\n- Don't change the top-level structure of existing robot-docs — add alongside existing fields","status":"open","priority":4,"issue_type":"feature","created_at":"2026-01-30T20:26:43.169688Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:32.090671Z","compaction_level":0,"original_size":0,"labels":["enhancement","robot-mode"]}
 {"id":"bd-2ac","title":"Create migration 009_embeddings.sql","description":"## Background\nMigration 009 creates the embedding storage layer for Gate B. It introduces a sqlite-vec vec0 virtual table for vector search and an embedding_metadata table for tracking provenance per chunk. Unlike migrations 007-008, this migration REQUIRES sqlite-vec to be loaded before it can be applied. The migration runner in db.rs must load the sqlite-vec extension first.\n\n## Approach\nCreate `migrations/009_embeddings.sql` per PRD Section 1.3.\n\n**Tables:**\n1. `embeddings` — vec0 virtual table with `embedding float[768]`\n2. `embedding_metadata` — tracks per-chunk provenance with composite PK (document_id, chunk_index)\n3. Orphan cleanup trigger: `documents_embeddings_ad` — deletes ALL chunk embeddings when a document is deleted using range deletion `[doc_id * 1000, (doc_id + 1) * 1000)`\n\n**Critical: sqlite-vec loading:**\nThe migration runner in `src/core/db.rs` must load sqlite-vec BEFORE applying any migrations. This means adding extension loading to the `create_connection()` or `run_migrations()` function. sqlite-vec is loaded via:\n```rust\nconn.load_extension_enable()?;\nconn.load_extension(\"vec0\", None)?;  // or platform-specific path\nconn.load_extension_disable()?;\n```\n\nRegister migration 9 in `src/core/db.rs` MIGRATIONS array.\n\n## Acceptance Criteria\n- [ ] `migrations/009_embeddings.sql` file exists\n- [ ] `embeddings` vec0 virtual table created with `embedding float[768]`\n- [ ] `embedding_metadata` table has composite PK (document_id, chunk_index)\n- [ ] `embedding_metadata.document_id` has FK to documents(id) ON DELETE CASCADE\n- [ ] Error tracking fields: last_error, attempt_count, last_attempt_at\n- [ ] Orphan cleanup trigger: deletes embeddings WHERE rowid in [doc_id*1000, (doc_id+1)*1000)\n- [ ] Index on embedding_metadata(last_error) WHERE last_error IS NOT NULL\n- [ ] Index on embedding_metadata(document_id)\n- [ ] Schema version 9 recorded\n- [ ] Migration runner loads sqlite-vec before applying migrations\n- [ ] `cargo build` succeeds\n\n## Files\n- `migrations/009_embeddings.sql` — new file (copy exact SQL from PRD Section 1.3)\n- `src/core/db.rs` — add migration 9 to MIGRATIONS array; add sqlite-vec extension loading\n\n## TDD Loop\nRED: Register migration in db.rs, `cargo test migration_tests` fails\nGREEN: Create SQL file + add extension loading\nVERIFY: `cargo test migration_tests && cargo build`\n\n## Edge Cases\n- sqlite-vec not installed: migration fails with clear error (not a silent skip)\n- Migration applied without sqlite-vec loaded: `CREATE VIRTUAL TABLE` fails with \"no such module: vec0\"\n- Documents deleted before embeddings: trigger fires but vec0 DELETE on empty range is safe\n- vec0 doesn't support FK cascades: that's why we need the explicit trigger","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:33.958178Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:22:26.478290Z","closed_at":"2026-01-30T17:22:26.478229Z","close_reason":"Completed: migration 009_embeddings.sql with vec0 table, embedding_metadata with composite PK, orphan cleanup trigger, registered in db.rs","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2ac","depends_on_id":"bd-221","type":"blocks","created_at":"2026-01-30T15:29:24.594861Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2as","title":"[CP1] Epic: Issue Ingestion","description":"Ingest all issues, labels, and issue discussions from configured GitLab repositories with resumable cursor-based incremental sync. This establishes the core data ingestion pattern reused for MRs in CP2.\n\nSuccess Criteria:\n- gi ingest --type=issues fetches all issues (count matches GitLab UI)\n- Labels extracted from issue payloads\n- Issue discussions fetched per-issue\n- Cursor-based sync is resumable\n- Sync tracking records all runs\n- Single-flight lock prevents concurrent runs\n\nReference: docs/prd/checkpoint-1.md","status":"tombstone","priority":1,"issue_type":"task","created_at":"2026-01-25T15:18:44.062057Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.155746Z","deleted_at":"2026-01-25T15:21:35.155744Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2bu","title":"[CP1] GitLab types for issues, discussions, notes","description":"Add Rust types to src/gitlab/types.rs for GitLab API responses.\n\n## Types to Add\n\n### GitLabIssue\n- id: i64 (GitLab global ID)\n- iid: i64 (project-scoped issue number)\n- project_id: i64\n- title: String\n- description: Option<String>\n- state: String (\"opened\" | \"closed\")\n- created_at, updated_at: String (ISO 8601)\n- closed_at: Option<String>\n- author: GitLabAuthor\n- labels: Vec<String> (array of label names - CP1 canonical)\n- web_url: String\nNOTE: labels_details intentionally NOT modeled - varies across GitLab versions\n\n### GitLabAuthor\n- id: i64\n- username: String\n- name: String\n\n### GitLabDiscussion\n- id: String (like \"6a9c1750b37d...\")\n- individual_note: bool\n- notes: Vec<GitLabNote>\n\n### GitLabNote\n- id: i64\n- note_type: Option<String> (\"DiscussionNote\" | \"DiffNote\" | null)\n- body: String\n- author: GitLabAuthor\n- created_at, updated_at: String (ISO 8601)\n- system: bool\n- resolvable: bool (default false)\n- resolved: bool (default false)\n- resolved_by: Option<GitLabAuthor>\n- resolved_at: Option<String>\n- position: Option<GitLabNotePosition>\n\n### GitLabNotePosition\n- old_path, new_path: Option<String>\n- old_line, new_line: Option<i32>\n\nFiles: src/gitlab/types.rs\nTests: Test deserialization with fixtures\nDone when: Types compile and deserialize sample API responses correctly","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:42:46.922805Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.710057Z","deleted_at":"2026-01-25T17:02:01.710051Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2cu","title":"[CP1] Discussion ingestion module","description":"Fetch and store discussions/notes for each issue.\n\n## Module\nsrc/ingestion/discussions.rs\n\n## Key Structs\n\n### IngestDiscussionsResult\n- discussions_fetched: usize\n- discussions_upserted: usize\n- notes_upserted: usize\n- system_notes_count: usize\n\n## Main Function\npub async fn ingest_issue_discussions(\n    conn, client, config,\n    project_id, gitlab_project_id,\n    issue_iid, local_issue_id, issue_updated_at\n) -> Result<IngestDiscussionsResult>\n\n## Logic\n1. Paginate through all discussions for given issue\n2. For each discussion:\n   - Begin transaction\n   - Store raw payload (compressed)\n   - Transform and upsert discussion record with correct issue FK\n   - Get local discussion ID\n   - Transform notes from discussion\n   - For each note:\n     - Store raw payload\n     - Upsert note with discussion_id FK\n     - Count system notes\n   - Commit transaction\n3. After all discussions synced: mark_discussions_synced(conn, local_issue_id, issue_updated_at)\n   - UPDATE issues SET discussions_synced_for_updated_at = ? WHERE id = ?\n\n## Invariant\nA rerun MUST NOT refetch discussions for issues whose updated_at has not advanced, even with cursor rewind. The discussions_synced_for_updated_at watermark ensures this.\n\n## Helper Functions\n- upsert_discussion(conn, discussion, payload_id)\n- get_local_discussion_id(conn, project_id, gitlab_id) -> i64\n- upsert_note(conn, discussion_id, note, payload_id)\n- mark_discussions_synced(conn, issue_id, issue_updated_at)\n\nFiles: src/ingestion/discussions.rs\nTests: tests/discussion_watermark_tests.rs\nDone when: Discussions and notes populated with correct FKs, watermark prevents redundant refetch","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T16:57:36.703237Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.827880Z","deleted_at":"2026-01-25T17:02:01.827876Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2dk","title":"Implement project resolution for --project filter","description":"## Background\nThe --project filter on search (and other commands) accepts a string that must be resolved to a project_id. Users may type the full path, a partial path, or just the project name. The resolution logic provides cascading match with helpful error messages when ambiguous. This improves ergonomics for multi-project installations.\n\n## Approach\nImplement project resolution function (location TBD — likely `src/core/project.rs` or inline in search filters):\n\n```rust\n/// Resolve a project string to a project_id using cascading match:\n/// 1. Exact match on path_with_namespace\n/// 2. Case-insensitive exact match\n/// 3. Suffix match (only if unambiguous)\n/// 4. Error with available projects list\npub fn resolve_project(conn: &Connection, project_str: &str) -> Result<i64>\n```\n\n**SQL queries:**\n```sql\n-- Step 1: exact match\nSELECT id FROM projects WHERE path_with_namespace = ?\n\n-- Step 2: case-insensitive\nSELECT id FROM projects WHERE LOWER(path_with_namespace) = LOWER(?)\n\n-- Step 3: suffix match\nSELECT id, path_with_namespace FROM projects\nWHERE path_with_namespace LIKE '%/' || ?\n   OR path_with_namespace = ?\n\n-- Step 4: list all for error message\nSELECT path_with_namespace FROM projects ORDER BY path_with_namespace\n```\n\n**Error format:**\n```\nError: Project 'auth-service' not found.\n\nAvailable projects:\n  backend/auth-service\n  frontend/auth-service-ui\n  infra/auth-proxy\n\nHint: Use the full path, e.g., --project=backend/auth-service\n```\n\nUses `LoreError::Ambiguous` variant for multiple suffix matches.\n\n## Acceptance Criteria\n- [ ] Exact match: \"group/project\" resolves correctly\n- [ ] Case-insensitive: \"Group/Project\" resolves to \"group/project\"\n- [ ] Suffix match: \"project-name\" resolves when only one \"*/project-name\" exists\n- [ ] Ambiguous suffix: error lists matching projects with hint\n- [ ] No match: error lists all available projects with hint\n- [ ] Empty projects table: clear error message\n- [ ] `cargo test project_resolution` passes\n\n## Files\n- `src/core/project.rs` — new file (or add to existing module)\n- `src/core/mod.rs` — add `pub mod project;`\n\n## TDD Loop\nRED: Tests:\n- `test_exact_match` — full path resolves\n- `test_case_insensitive` — mixed case resolves\n- `test_suffix_unambiguous` — short name resolves when unique\n- `test_suffix_ambiguous` — error with list when multiple match\n- `test_no_match` — error with available projects\nGREEN: Implement resolve_project\nVERIFY: `cargo test project_resolution`\n\n## Edge Cases\n- Project path containing special LIKE characters (%, _): unlikely but escape for safety\n- Single project in DB: suffix always unambiguous\n- Project path with multiple slashes: \"org/group/project\" — suffix match on \"project\" works","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:26:13.076571Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:39:17.197735Z","closed_at":"2026-01-30T17:39:17.197552Z","close_reason":"Implemented resolve_project() with cascading match (exact, CI, suffix) + 6 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2dk","depends_on_id":"bd-3q2","type":"blocks","created_at":"2026-01-30T15:29:24.446650Z","created_by":"tayloreernisse"}]}
+{"id":"bd-2e8","title":"Add fetchResourceEvents config flag to SyncConfig","description":"## Background\nEvent fetching should be opt-in (default true) so users who don't need temporal queries skip 3 extra API calls per entity. This follows the existing SyncConfig pattern with serde defaults and camelCase JSON aliases.\n\n## Approach\nAdd to SyncConfig in src/core/config.rs:\n```rust\n#[serde(rename = \"fetchResourceEvents\", default = \"default_true\")]\npub fetch_resource_events: bool,\n```\n\nAdd default function (if not already present):\n```rust\nfn default_true() -> bool { true }\n```\n\nUpdate Default impl for SyncConfig to include `fetch_resource_events: true`.\n\nAdd --no-events flag to sync command in src/cli/mod.rs (SyncArgs):\n```rust\n/// Skip resource event fetching (overrides config)\n#[arg(long = \"no-events\", help_heading = \"Sync Options\")]\npub no_events: bool,\n```\n\nIn the sync command handler (src/cli/commands/sync.rs), override config when flag is set:\n```rust\nif args.no_events {\n    config.sync.fetch_resource_events = false;\n}\n```\n\n## Acceptance Criteria\n- [ ] SyncConfig deserializes `fetchResourceEvents: false` from JSON config\n- [ ] SyncConfig defaults to `fetch_resource_events: true` when field absent\n- [ ] `--no-events` flag parses correctly in CLI\n- [ ] `--no-events` overrides config to false\n- [ ] `cargo test` passes with no regressions\n\n## Files\n- src/core/config.rs (add field to SyncConfig + default fn + Default impl)\n- src/cli/mod.rs (add --no-events to SyncArgs)\n- src/cli/commands/sync.rs (override config when flag set)\n\n## TDD Loop\nRED: tests/config_tests.rs (or inline in config.rs):\n- `test_sync_config_fetch_resource_events_default_true` - omit field from JSON, verify default\n- `test_sync_config_fetch_resource_events_explicit_false` - set field false, verify parsed\n- `test_sync_config_no_events_flag` - verify CLI arg parsing\n\nGREEN: Add the field, default fn, Default impl update, CLI flag, and override logic\n\nVERIFY: `cargo test config -- --nocapture && cargo build`\n\n## Edge Cases\n- Ensure serde rename matches camelCase convention used by all other SyncConfig fields\n- The default_true fn may already exist for other fields — check before adding duplicate\n- The --no-events flag must NOT be confused with --no-X negation flags already in CLI (check mod.rs for conflicts)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:24.006037Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:41:12.107758Z","compaction_level":0,"original_size":0,"labels":["config","gate-1","phase-b"],"dependencies":[{"issue_id":"bd-2e8","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:24.010608Z","created_by":"tayloreernisse"}]}
+{"id":"bd-2ez","title":"Add 'lore count references' command","description":"## Background\nThe entity_references table is the cross-reference graph powering Gates 3-5. Operators and agents need a way to inspect this graph for debugging and validation: how many refs exist, breakdown by type and method, unresolved count. Follows the existing `lore count` pattern.\n\n## Approach\n1. Add 'references' to CountArgs entity value_parser in src/cli/mod.rs\n2. Add handler in src/cli/commands/count.rs matching the existing pattern:\n\n```rust\n\"references\" => {\n    let total: i64 = conn.query_row(\n        \"SELECT COUNT(*) FROM entity_references\", [], |r| r.get(0)\n    )?;\n    let by_type: Vec<(String, i64)> = conn.prepare(\n        \"SELECT reference_type, COUNT(*) FROM entity_references GROUP BY reference_type\"\n    )?.query_map([], |r| Ok((r.get(0)?, r.get(1)?)))?.collect::<Result<_,_>>()?;\n    let by_method: Vec<(String, i64)> = conn.prepare(\n        \"SELECT source_method, COUNT(*) FROM entity_references GROUP BY source_method\"\n    )?.query_map([], |r| Ok((r.get(0)?, r.get(1)?)))?.collect::<Result<_,_>>()?;\n    let unresolved: i64 = conn.query_row(\n        \"SELECT COUNT(*) FROM entity_references WHERE target_entity_id IS NULL\", [], |r| r.get(0)\n    )?;\n    // Format as table (human) or JSON (robot)\n}\n```\n\nRobot JSON:\n```json\n{\"ok\":true,\"data\":{\"total\":142,\"by_type\":{\"closes\":45,\"mentioned\":89,\"related\":8},\"by_method\":{\"api_closes_issues\":40,\"api_state_event\":5,\"system_note_parse\":97},\"unresolved\":12}}\n```\n\n## Acceptance Criteria\n- [ ] `lore count references` prints total, by-type, by-method, unresolved counts\n- [ ] `lore --robot count references` returns structured JSON\n- [ ] Human output uses the same table formatting as other count subcommands\n- [ ] Works with empty table (all zeros)\n- [ ] Exit code 0 on success\n\n## Files\n- src/cli/mod.rs (add 'references' to CountArgs value_parser)\n- src/cli/commands/count.rs (add references handler)\n\n## TDD Loop\nRED: tests/count_references_test.rs:\n- `test_count_references_empty` - no refs → all zeros\n- `test_count_references_with_data` - seed refs, verify counts match\n- `test_count_references_robot_json` - verify JSON structure\n\nGREEN: Add the count handler with SQL queries\nREFACTOR: Share query helpers if pattern is reusable\n\nVERIFY: `cargo test count_references -- --nocapture`\n\n## Edge Cases\n- entity_references table doesn't exist (pre-migration 011): return meaningful error\n- All references are unresolved: unresolved count equals total\n- Reference types not in CHECK constraint (future extension): GROUP BY still works","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T22:42:43.780303Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:43:40.652577Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2ez","depends_on_id":"bd-1se","type":"parent-child","created_at":"2026-02-02T22:43:40.652558Z","created_by":"tayloreernisse"},{"issue_id":"bd-2ez","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T22:43:33.877742Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2f0","title":"[CP1] gi count issues/discussions/notes commands","description":"## Background\n\nThe `gi count` command provides quick counts of entities in the local database. It supports counting issues, MRs, discussions, and notes, with optional filtering by noteable type. This enables quick validation that sync is working correctly.\n\n## Approach\n\n### Module: src/cli/commands/count.rs\n\n### Clap Definition\n\n```rust\n#[derive(Args)]\npub struct CountArgs {\n    /// Entity type to count\n    #[arg(value_parser = [\"issues\", \"mrs\", \"discussions\", \"notes\"])]\n    pub entity: String,\n\n    /// Filter by noteable type (for discussions/notes)\n    #[arg(long, value_parser = [\"issue\", \"mr\"])]\n    pub r#type: Option<String>,\n}\n```\n\n### Handler Function\n\n```rust\npub async fn handle_count(args: CountArgs, conn: &Connection) -> Result<()>\n```\n\n### Queries by Entity\n\n**issues:**\n```sql\nSELECT COUNT(*) FROM issues\n```\nOutput: `Issues: 3,801`\n\n**discussions:**\n```sql\n-- Without type filter\nSELECT COUNT(*) FROM discussions\n\n-- With --type=issue\nSELECT COUNT(*) FROM discussions WHERE noteable_type = 'Issue'\n```\nOutput: `Issue Discussions: 1,234`\n\n**notes:**\n```sql\n-- Total and system count\nSELECT COUNT(*), SUM(is_system) FROM notes\n\n-- With --type=issue (join through discussions)\nSELECT COUNT(*), SUM(n.is_system)\nFROM notes n\nJOIN discussions d ON n.discussion_id = d.id\nWHERE d.noteable_type = 'Issue'\n```\nOutput: `Issue Notes: 5,678 (excluding 1,234 system)`\n\n### Output Format\n\n```\nIssues: 3,801\n```\n\n```\nIssue Discussions: 1,234\n```\n\n```\nIssue Notes: 5,678 (excluding 1,234 system)\n```\n\n## Acceptance Criteria\n\n- [ ] `gi count issues` shows total issue count\n- [ ] `gi count discussions` shows total discussion count\n- [ ] `gi count discussions --type=issue` filters to issue discussions\n- [ ] `gi count notes` shows total note count with system note exclusion\n- [ ] `gi count notes --type=issue` filters to issue notes\n- [ ] Numbers formatted with thousands separators (1,234)\n\n## Files\n\n- src/cli/commands/mod.rs (add `pub mod count;`)\n- src/cli/commands/count.rs (create)\n- src/cli/mod.rs (add Count variant to Commands enum)\n\n## TDD Loop\n\nRED:\n```rust\n#[tokio::test] async fn count_issues_returns_total()\n#[tokio::test] async fn count_discussions_with_type_filter()\n#[tokio::test] async fn count_notes_excludes_system_notes()\n```\n\nGREEN: Implement handler with queries\n\nVERIFY: `cargo test count`\n\n## Edge Cases\n\n- Zero entities - show \"Issues: 0\"\n- --type flag invalid for issues/mrs - ignore or error\n- All notes are system notes - show \"Notes: 0 (excluding 1,234 system)\"","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-25T17:02:38.360495Z","created_by":"tayloreernisse","updated_at":"2026-01-25T23:01:37.084627Z","closed_at":"2026-01-25T23:01:37.084568Z","close_reason":"Implemented gi count command with issues/discussions/notes support, format_number helper, and system note exclusion","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2f0","depends_on_id":"bd-208","type":"blocks","created_at":"2026-01-25T17:04:05.677181Z","created_by":"tayloreernisse"}]}
+{"id":"bd-2f2","title":"Implement timeline human output renderer","description":"## Background\nHuman output for timeline needs colored, readable format per spec §3.4. Uses the existing color infrastructure (--color flag support from src/cli/mod.rs). The output format is a vertical timeline with date, event type, entity ref, summary, and actor.\n\n## Approach\nCreate src/cli/commands/timeline.rs (new file):\n\n```rust\nuse crate::core::timeline::{TimelineResult, TimelineEvent, TimelineEventType};\n\n/// Print timeline in human-readable format with optional colors.\npub fn print_timeline(result: &TimelineResult, use_color: bool) {\n    // Header\n    println!(\"Timeline: \\\"{}\\\" ({} events across {} entities)\",\n        result.query, result.events.len(),\n        result.seed_entities.len() + result.expanded_entities.len());\n    println!(\"{}\", \"─\".repeat(55));\n    println!();\n\n    // Events\n    for event in &result.events {\n        let date = format_date(event.timestamp);  // YYYY-MM-DD\n        let event_tag = format_event_tag(&event.event_type, use_color);  // CREATED, LABEL, etc.\n        let entity_ref = format_entity_ref(event);  // #234 or !567\n        let expanded_marker = if !event.is_seed { \" [expanded]\" } else { \"\" };\n        \n        println!(\"{:10}  {:8}  {:5}  {:<40}  @{}{}\",\n            date, event_tag, entity_ref, event.summary,\n            event.actor.as_deref().unwrap_or(\"-\"), expanded_marker);\n        \n        // Evidence notes: show snippet indented in quotes\n        if let TimelineEventType::NoteEvidence { snippet, .. } = &event.event_type {\n            println!(\"                             \\\"{}\\\"\", snippet);\n        }\n    }\n\n    // Footer\n    println!();\n    println!(\"{}\", \"─\".repeat(55));\n    print!(\"Seed entities: \");\n    // list seed entity refs\n    if !result.expanded_entities.is_empty() {\n        print!(\" | Expanded: \");\n        // list expanded with depth + via info\n    }\n    println!();\n}\n```\n\n**Event tag formatting:**\n- Created → \"CREATED\" (green)\n- StateChanged(\"closed\") → \"CLOSED\" (red)\n- StateChanged(\"reopened\") → \"REOPENED\" (yellow)\n- StateChanged(\"merged\") → \"MERGED\" (magenta)\n- LabelAdded → \"LABEL+\" (cyan)\n- LabelRemoved → \"LABEL-\" (cyan)\n- MilestoneSet → \"MILESTONE\" (blue)\n- Merged → \"MERGED\" (magenta)\n- NoteEvidence → \"NOTE\" (white/default)\n- CrossReferenced → \"XREF\" (yellow)\n\n**Entity ref formatting:**\n- issue → \"#\" + iid\n- merge_request → \"!\" + iid\n\n**Expanded entity annotation:**\n```\n2024-03-28  CREATED   #299  OAuth2 login fails for SSO users     @dave   [expanded]\n            (via !567, closes)\n```\n\n## Acceptance Criteria\n- [ ] Header shows query, event count, entity count\n- [ ] Each event shows date, tag, entity ref, summary, actor\n- [ ] Evidence notes show snippet in quotes, indented\n- [ ] Expanded entities marked with [expanded] and provenance\n- [ ] Footer lists seed and expanded entities\n- [ ] Colors applied when use_color=true\n- [ ] No colors when use_color=false (piped output)\n\n## Files\n- src/cli/commands/timeline.rs (new)\n- src/cli/commands/mod.rs (add `pub mod timeline;`)\n\n## TDD Loop\nRED: tests/timeline_output_tests.rs:\n- `test_print_timeline_header` - capture stdout, verify header format\n- `test_print_timeline_event_formatting` - verify date/tag/ref/summary layout\n- `test_print_timeline_evidence_note_indented` - verify snippet in quotes\n- `test_print_timeline_expanded_marker` - verify [expanded] and provenance\n- `test_print_timeline_no_color` - verify no ANSI codes when color off\n\nNote: Capturing stdout in tests requires a buffer-based approach. Consider making print_timeline write to impl Write instead of stdout directly for testability.\n\nGREEN: Implement the human output renderer\n\nVERIFY: `cargo test timeline_output -- --nocapture`\n\n## Edge Cases\n- Very long summaries: truncate to ~50 chars for table alignment\n- Actors with long usernames: alignment may shift — use fixed-width columns\n- Empty timeline (no events): print \"No events found\" message\n- Timestamp conversion: ms epoch → YYYY-MM-DD display, handle timezone (UTC)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:28.326026Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:46:58.820468Z","compaction_level":0,"original_size":0,"labels":["cli","gate-3","phase-b"],"dependencies":[{"issue_id":"bd-2f2","depends_on_id":"bd-3as","type":"blocks","created_at":"2026-02-02T21:33:37.659719Z","created_by":"tayloreernisse"},{"issue_id":"bd-2f2","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:28.329132Z","created_by":"tayloreernisse"}]}
+{"id":"bd-2fc","title":"Update AGENTS.md and CLAUDE.md with Phase B commands","description":"## Background\nAGENTS.md and the user's CLAUDE.md contain robot-mode documentation for agent consumption. After Phase B ships, the new commands (timeline, file-history, trace) and the extended count command must be documented so agents know they exist and how to use them.\n\n## Approach\nAdd a new section to AGENTS.md under '### Phase B: Temporal Intelligence Commands':\n\n```markdown\n### Temporal Intelligence Commands\n\n```bash\n# Entity decision timeline\nlore --robot timeline issue 123\nlore --robot timeline mr 456 --since 30d\nlore --robot timeline issue 123 --type state,label\n\n# File decision history\nlore --robot file-history src/core/db.rs\nlore --robot file-history src/core/db.rs --since 90d\n\n# Code trace (file → MRs → issues → cross-refs)\nlore --robot trace src/core/db.rs\nlore --robot trace 'src/embedding/**' --depth 2\n\n# Cross-reference stats\nlore --robot count references\n```\n```\n\nAlso update the Commands table in AGENTS.md Robot Mode section to include the new commands.\n\nFor CLAUDE.md: the user maintains this separately at ~/.claude/CLAUDE.md. Add a note in AGENTS.md that agents should check CLAUDE.md for the latest command list.\n\n## Acceptance Criteria\n- [ ] AGENTS.md has a 'Temporal Intelligence Commands' section\n- [ ] All 4 new commands documented with examples\n- [ ] Commands table updated with timeline, file-history, trace entries\n- [ ] Examples use --robot flag consistently\n- [ ] Documentation matches actual CLI behavior (flags, args, output format)\n\n## Files\n- AGENTS.md (update existing robot mode section)\n\n## TDD Loop\nThis is a documentation bead — no automated tests. Verify manually:\n- [ ] Read AGENTS.md and confirm new section exists\n- [ ] Verify each example command matches CLI arg definitions in src/cli/mod.rs\n- [ ] Run `lore --robot timeline --help` and confirm flags match docs\n\n## Edge Cases\n- Don't document commands before they're implemented — this bead should be one of the last in the dependency chain\n- Keep examples consistent with existing documentation style (no emojis, concise)","status":"open","priority":4,"issue_type":"task","created_at":"2026-02-02T22:43:22.090741Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:43:40.829868Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2fc","depends_on_id":"bd-1ht","type":"parent-child","created_at":"2026-02-02T22:43:40.829848Z","created_by":"tayloreernisse"},{"issue_id":"bd-2fc","depends_on_id":"bd-1v8","type":"blocks","created_at":"2026-02-02T22:43:34.047898Z","created_by":"tayloreernisse"}]}
+{"id":"bd-2fm","title":"Add GitLab Resource Event serde types","description":"## Background\nNeed Rust types for deserializing GitLab Resource Events API responses. These map directly to the API JSON shape from three endpoints: resource_state_events, resource_label_events, resource_milestone_events.\n\nExisting pattern: types.rs uses #[derive(Debug, Clone, Deserialize)] with Option<T> for nullable fields. GitLabAuthor is already defined (id, username, name). Tests in tests/gitlab_types_tests.rs use serde_json::from_str with sample payloads.\n\n## Approach\nAdd to src/gitlab/types.rs (after existing types):\n\n```rust\n/// Reference to an MR in state event's source_merge_request field\n#[derive(Debug, Clone, Deserialize, Serialize)]\npub struct GitLabMergeRequestRef {\n    pub iid: i64,\n    pub title: Option<String>,\n    pub web_url: Option<String>,\n}\n\n/// Reference to a label in label event's label field\n#[derive(Debug, Clone, Deserialize, Serialize)]\npub struct GitLabLabelRef {\n    pub id: i64,\n    pub name: String,\n    pub color: Option<String>,\n    pub description: Option<String>,\n}\n\n/// Reference to a milestone in milestone event's milestone field\n#[derive(Debug, Clone, Deserialize, Serialize)]\npub struct GitLabMilestoneRef {\n    pub id: i64,\n    pub iid: i64,\n    pub title: String,\n}\n\n#[derive(Debug, Clone, Deserialize, Serialize)]\npub struct GitLabStateEvent {\n    pub id: i64,\n    pub user: Option<GitLabAuthor>,\n    pub created_at: String,\n    pub resource_type: String,       // \"Issue\" | \"MergeRequest\"\n    pub resource_id: i64,\n    pub state: String,               // \"opened\" | \"closed\" | \"reopened\" | \"merged\" | \"locked\"\n    pub source_commit: Option<String>,\n    pub source_merge_request: Option<GitLabMergeRequestRef>,\n}\n\n#[derive(Debug, Clone, Deserialize, Serialize)]\npub struct GitLabLabelEvent {\n    pub id: i64,\n    pub user: Option<GitLabAuthor>,\n    pub created_at: String,\n    pub resource_type: String,\n    pub resource_id: i64,\n    pub label: GitLabLabelRef,\n    pub action: String,              // \"add\" | \"remove\"\n}\n\n#[derive(Debug, Clone, Deserialize, Serialize)]\npub struct GitLabMilestoneEvent {\n    pub id: i64,\n    pub user: Option<GitLabAuthor>,\n    pub created_at: String,\n    pub resource_type: String,\n    pub resource_id: i64,\n    pub milestone: GitLabMilestoneRef,\n    pub action: String,              // \"add\" | \"remove\"\n}\n```\n\nAlso export from src/gitlab/mod.rs if needed.\n\n## Acceptance Criteria\n- [ ] All 6 types (3 events + 3 refs) compile\n- [ ] GitLabStateEvent deserializes from real GitLab API JSON (with and without source_merge_request)\n- [ ] GitLabLabelEvent deserializes with nested label object\n- [ ] GitLabMilestoneEvent deserializes with nested milestone object\n- [ ] All Optional fields handle null/missing correctly\n- [ ] Types exported from lore::gitlab::types\n\n## Files\n- src/gitlab/types.rs (add 6 new types)\n- tests/gitlab_types_tests.rs (add deserialization tests)\n\n## TDD Loop\nRED: Add to tests/gitlab_types_tests.rs:\n- `test_deserialize_state_event_closed_by_mr` - JSON with source_merge_request present\n- `test_deserialize_state_event_simple` - JSON with source_merge_request null, user null\n- `test_deserialize_label_event_add` - label add with full label object\n- `test_deserialize_label_event_remove` - label remove\n- `test_deserialize_milestone_event` - milestone add with nested milestone\nImport new types: `use lore::gitlab::types::{GitLabStateEvent, GitLabLabelEvent, GitLabMilestoneEvent, GitLabMergeRequestRef, GitLabLabelRef, GitLabMilestoneRef};`\n\nGREEN: Add the type definitions to types.rs\n\nVERIFY: `cargo test gitlab_types_tests -- --nocapture`\n\n## Edge Cases\n- GitLab sometimes returns user: null for system-generated events (e.g., auto-close on merge) — user must be Option\n- source_merge_request can be null even when state is \"closed\" (manually closed, not by MR)\n- label.color may be null for labels created via API without color\n- The resource_type field uses PascalCase (\"MergeRequest\" not \"merge_request\") — don't confuse with DB entity_type","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:24.081234Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:41:31.022308Z","compaction_level":0,"original_size":0,"labels":["gate-1","phase-b","types"],"dependencies":[{"issue_id":"bd-2fm","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:24.085809Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2fp","title":"Implement discussion document extraction","description":"## Background\nDiscussion documents are the most complex extraction — they involve querying discussions + notes + parent entity (issue or MR) + parent labels + DiffNote file paths. The output includes a threaded conversation format with author/date prefixes per note. System notes (bot-generated) are excluded. DiffNote paths are extracted for the --path search filter.\n\n## Approach\nImplement `extract_discussion_document()` in `src/documents/extractor.rs`:\n\n```rust\n/// Extract a searchable document from a discussion thread.\n/// Returns None if the discussion or its parent has been deleted.\npub fn extract_discussion_document(conn: &Connection, discussion_id: i64) -> Result<Option<DocumentData>>\n```\n\n**SQL queries (from PRD Section 2.2):**\n```sql\n-- Discussion metadata\nSELECT d.id, d.noteable_type, d.issue_id, d.merge_request_id,\n       p.path_with_namespace, p.id AS project_id\nFROM discussions d\nJOIN projects p ON p.id = d.project_id\nWHERE d.id = ?\n\n-- Parent entity (conditional on noteable_type)\n-- If Issue: SELECT i.iid, i.title, i.web_url FROM issues i WHERE i.id = ?\n-- If MR: SELECT m.iid, m.title, m.web_url FROM merge_requests m WHERE m.id = ?\n\n-- Parent labels (via issue_labels or mr_labels junction)\n\n-- Non-system notes in thread order\nSELECT n.author_username, n.body, n.created_at, n.gitlab_id,\n       n.note_type, n.position_old_path, n.position_new_path\nFROM notes n\nWHERE n.discussion_id = ? AND n.is_system = 0\nORDER BY n.created_at ASC, n.id ASC\n```\n\n**Document format:**\n```\n[[Discussion]] Issue #234: Authentication redesign\nProject: group/project-one\nURL: https://gitlab.example.com/group/project-one/-/issues/234#note_12345\nLabels: [\"bug\", \"auth\"]\nFiles: [\"src/auth/login.ts\"]\n\n--- Thread ---\n\n@johndoe (2024-03-15):\nI think we should move to JWT-based auth...\n\n@janedoe (2024-03-15):\nAgreed. What about refresh token strategy?\n```\n\n**Implementation steps:**\n1. Query discussion row — if not found, return Ok(None)\n2. Determine parent type (Issue or MR) from noteable_type\n3. Query parent entity for iid, title, web_url — if not found, return Ok(None)\n4. Query parent labels via appropriate junction table\n5. Query non-system notes ordered by created_at ASC, id ASC\n6. Extract DiffNote paths: collect position_old_path and position_new_path, dedup\n7. Construct URL: `{parent_web_url}#note_{first_note_gitlab_id}`\n8. Format header with [[Discussion]] prefix\n9. Format thread body: `@author (YYYY-MM-DD):\\nbody\\n\\n` per note\n10. Apply discussion truncation via `truncate_discussion()` if needed\n11. Author = first non-system note's author_username\n12. Compute hashes, return DocumentData\n\n## Acceptance Criteria\n- [ ] System notes (is_system=1) excluded from content\n- [ ] DiffNote paths extracted from position_old_path and position_new_path\n- [ ] Paths deduplicated and sorted\n- [ ] URL constructed as `parent_web_url#note_GITLAB_ID`\n- [ ] Header uses parent entity type: \"Issue #N\" or \"MR !N\"\n- [ ] Parent title included in header\n- [ ] Labels come from PARENT entity (not the discussion itself)\n- [ ] First non-system note author used as document author\n- [ ] Thread formatted with `@author (date):` per note\n- [ ] Truncation applied for long threads via truncate_discussion()\n- [ ] `cargo test extract_discussion` passes\n\n## Files\n- `src/documents/extractor.rs` — implement `extract_discussion_document()`\n\n## TDD Loop\nRED: Tests in `#[cfg(test)] mod tests`:\n- `test_discussion_document_format` — verify header + thread format\n- `test_discussion_not_found` — returns Ok(None)\n- `test_discussion_parent_deleted` — returns Ok(None) when parent issue/MR missing\n- `test_discussion_system_notes_excluded` — system notes not in content\n- `test_discussion_diffnote_paths` — old_path + new_path extracted and deduped\n- `test_discussion_url_construction` — URL has #note_GITLAB_ID anchor\n- `test_discussion_uses_parent_labels` — labels from parent entity, not discussion\nGREEN: Implement extract_discussion_document\nVERIFY: `cargo test extract_discussion`\n\n## Edge Cases\n- Discussion with all system notes: no non-system notes -> return empty thread (or skip document entirely?)\n- Discussion with NULL parent (orphaned): return Ok(None)\n- DiffNote with same old_path and new_path: dedup produces single entry\n- Notes with NULL body: skip or use empty string\n- Discussion on MR: header shows \"MR !N\" (not \"MergeRequest !N\")","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:25:45.549099Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:34:43.597398Z","closed_at":"2026-01-30T17:34:43.597339Z","close_reason":"Implemented extract_discussion_document() with parent entity lookup, DiffNote paths, system note exclusion, URL construction + 9 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2fp","depends_on_id":"bd-18t","type":"blocks","created_at":"2026-01-30T15:29:15.914098Z","created_by":"tayloreernisse"},{"issue_id":"bd-2fp","depends_on_id":"bd-36p","type":"blocks","created_at":"2026-01-30T15:29:15.847680Z","created_by":"tayloreernisse"},{"issue_id":"bd-2fp","depends_on_id":"bd-hrs","type":"blocks","created_at":"2026-01-30T15:29:15.880008Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2h0","title":"[CP1] gi list issues command","description":"List issues from the database.\n\n## Module\nsrc/cli/commands/list.rs\n\n## Clap Definition\nList {\n    #[arg(value_parser = [\"issues\", \"mrs\"])]\n    entity: String,\n    \n    #[arg(long, default_value = \"20\")]\n    limit: usize,\n    \n    #[arg(long)]\n    project: Option<String>,\n    \n    #[arg(long, value_parser = [\"opened\", \"closed\", \"all\"])]\n    state: Option<String>,\n}\n\n## Output Format\nIssues (showing 20 of 3,801)\n\n  #1234  Authentication redesign                       opened   @johndoe   3 days ago\n  #1233  Fix memory leak in cache                      closed   @janedoe   5 days ago\n  #1232  Add dark mode support                         opened   @bobsmith  1 week ago\n  ...\n\n## Implementation\n- Query issues table with filters\n- Join with projects table for display\n- Format updated_at as relative time (\"3 days ago\")\n- Truncate title if too long\n\nFiles: src/cli/commands/list.rs\nDone when: List displays issues with proper filtering and formatting","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:58:23.809829Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.898106Z","deleted_at":"2026-01-25T17:02:01.898102Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2iq","title":"[CP1] Database migration 002_issues.sql","description":"## Background\n\nThe 002_issues.sql migration creates tables for issues, labels, issue_labels, discussions, and notes. This is the data foundation for Checkpoint 1, enabling issue ingestion with cursor-based sync, label tracking, and discussion storage.\n\n## Approach\n\nCreate `migrations/002_issues.sql` with complete SQL statements.\n\n### Full Migration SQL\n\n```sql\n-- Migration 002: Issue Ingestion Tables\n-- Applies on top of 001_initial.sql\n\n-- Issues table\nCREATE TABLE issues (\n    id INTEGER PRIMARY KEY,\n    gitlab_id INTEGER UNIQUE NOT NULL,\n    project_id INTEGER NOT NULL REFERENCES projects(id) ON DELETE CASCADE,\n    iid INTEGER NOT NULL,\n    title TEXT,\n    description TEXT,\n    state TEXT NOT NULL CHECK (state IN ('opened', 'closed')),\n    author_username TEXT,\n    created_at INTEGER NOT NULL,           -- ms epoch UTC\n    updated_at INTEGER NOT NULL,           -- ms epoch UTC\n    last_seen_at INTEGER NOT NULL,         -- updated on every upsert\n    discussions_synced_for_updated_at INTEGER, -- watermark for dependent sync\n    web_url TEXT,\n    raw_payload_id INTEGER REFERENCES raw_payloads(id)\n);\n\nCREATE INDEX idx_issues_project_updated ON issues(project_id, updated_at);\nCREATE INDEX idx_issues_author ON issues(author_username);\nCREATE UNIQUE INDEX uq_issues_project_iid ON issues(project_id, iid);\n\n-- Labels table (name-only for CP1)\nCREATE TABLE labels (\n    id INTEGER PRIMARY KEY,\n    gitlab_id INTEGER,                      -- optional, for future Labels API\n    project_id INTEGER NOT NULL REFERENCES projects(id) ON DELETE CASCADE,\n    name TEXT NOT NULL,\n    color TEXT,\n    description TEXT\n);\n\nCREATE UNIQUE INDEX uq_labels_project_name ON labels(project_id, name);\nCREATE INDEX idx_labels_name ON labels(name);\n\n-- Issue-label junction (DELETE before INSERT for stale removal)\nCREATE TABLE issue_labels (\n    issue_id INTEGER NOT NULL REFERENCES issues(id) ON DELETE CASCADE,\n    label_id INTEGER NOT NULL REFERENCES labels(id) ON DELETE CASCADE,\n    PRIMARY KEY(issue_id, label_id)\n);\n\nCREATE INDEX idx_issue_labels_label ON issue_labels(label_id);\n\n-- Discussion threads for issues (MR discussions added in CP2)\nCREATE TABLE discussions (\n    id INTEGER PRIMARY KEY,\n    gitlab_discussion_id TEXT NOT NULL,     -- GitLab string ID (e.g., \"6a9c1750b37d...\")\n    project_id INTEGER NOT NULL REFERENCES projects(id) ON DELETE CASCADE,\n    issue_id INTEGER REFERENCES issues(id) ON DELETE CASCADE,\n    merge_request_id INTEGER,               -- FK added in CP2 via ALTER TABLE\n    noteable_type TEXT NOT NULL CHECK (noteable_type IN ('Issue', 'MergeRequest')),\n    individual_note INTEGER NOT NULL DEFAULT 0,  -- 0=threaded, 1=standalone\n    first_note_at INTEGER,                  -- min(note.created_at) for ordering\n    last_note_at INTEGER,                   -- max(note.created_at) for \"recently active\"\n    last_seen_at INTEGER NOT NULL,          -- updated on every upsert\n    resolvable INTEGER NOT NULL DEFAULT 0,  -- MR discussions can be resolved\n    resolved INTEGER NOT NULL DEFAULT 0,\n    CHECK (\n        (noteable_type = 'Issue' AND issue_id IS NOT NULL AND merge_request_id IS NULL) OR\n        (noteable_type = 'MergeRequest' AND merge_request_id IS NOT NULL AND issue_id IS NULL)\n    )\n);\n\nCREATE UNIQUE INDEX uq_discussions_project_discussion_id ON discussions(project_id, gitlab_discussion_id);\nCREATE INDEX idx_discussions_issue ON discussions(issue_id);\nCREATE INDEX idx_discussions_mr ON discussions(merge_request_id);\nCREATE INDEX idx_discussions_last_note ON discussions(last_note_at);\n\n-- Notes belong to discussions\nCREATE TABLE notes (\n    id INTEGER PRIMARY KEY,\n    gitlab_id INTEGER UNIQUE NOT NULL,\n    discussion_id INTEGER NOT NULL REFERENCES discussions(id) ON DELETE CASCADE,\n    project_id INTEGER NOT NULL REFERENCES projects(id) ON DELETE CASCADE,\n    note_type TEXT,                         -- 'DiscussionNote' | 'DiffNote' | null\n    is_system INTEGER NOT NULL DEFAULT 0,   -- 1 for system-generated notes\n    author_username TEXT,\n    body TEXT,\n    created_at INTEGER NOT NULL,            -- ms epoch\n    updated_at INTEGER NOT NULL,            -- ms epoch\n    last_seen_at INTEGER NOT NULL,          -- updated on every upsert\n    position INTEGER,                       -- 0-indexed array order from API\n    resolvable INTEGER NOT NULL DEFAULT 0,\n    resolved INTEGER NOT NULL DEFAULT 0,\n    resolved_by TEXT,\n    resolved_at INTEGER,\n    -- DiffNote position metadata (populated for MR DiffNotes in CP2)\n    position_old_path TEXT,\n    position_new_path TEXT,\n    position_old_line INTEGER,\n    position_new_line INTEGER,\n    raw_payload_id INTEGER REFERENCES raw_payloads(id)\n);\n\nCREATE INDEX idx_notes_discussion ON notes(discussion_id);\nCREATE INDEX idx_notes_author ON notes(author_username);\nCREATE INDEX idx_notes_system ON notes(is_system);\n\n-- Update schema version\nINSERT INTO schema_version (version, applied_at, description)\nVALUES (2, strftime('%s', 'now') * 1000, 'Issue ingestion tables');\n```\n\n## Acceptance Criteria\n\n- [ ] Migration file exists at `migrations/002_issues.sql`\n- [ ] All tables created: issues, labels, issue_labels, discussions, notes\n- [ ] All indexes created as specified\n- [ ] CHECK constraints on state and noteable_type work correctly\n- [ ] CASCADE deletes work (project deletion cascades)\n- [ ] Migration applies cleanly on fresh DB after 001_initial.sql\n- [ ] schema_version updated to 2 after migration\n- [ ] `gi doctor` shows schema_version = 2\n\n## Files\n\n- migrations/002_issues.sql (create)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/migration_tests.rs\n#[test] fn migration_002_creates_issues_table()\n#[test] fn migration_002_creates_labels_table()\n#[test] fn migration_002_creates_discussions_table()\n#[test] fn migration_002_creates_notes_table()\n#[test] fn migration_002_enforces_state_check()\n#[test] fn migration_002_enforces_noteable_type_check()\n#[test] fn migration_002_cascades_on_project_delete()\n```\n\nGREEN: Create migration file with all SQL\n\nVERIFY:\n```bash\n# Apply migration to test DB\nsqlite3 :memory: < migrations/001_initial.sql\nsqlite3 :memory: < migrations/002_issues.sql\n\n# Verify schema_version\nsqlite3 test.db \"SELECT version FROM schema_version ORDER BY version DESC LIMIT 1\"\n# Expected: 2\n\ncargo test migration_002\n```\n\n## Edge Cases\n\n- Applying twice - should fail on UNIQUE constraint (idempotency via version check)\n- Missing 001 - foreign key to projects fails\n- Long label names - TEXT handles any length\n- NULL description - allowed by schema\n- Empty discussions_synced_for_updated_at - NULL means never synced","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.128594Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:25:10.309900Z","closed_at":"2026-01-25T22:25:10.309852Z","close_reason":"Created 002_issues.sql with issues/labels/issue_labels/discussions/notes tables, 8 passing tests verify schema, constraints, and cascades","compaction_level":0,"original_size":0}
 {"id":"bd-2ms","title":"[CP1] Unit tests for transformers","description":"Comprehensive unit tests for issue and discussion transformers.\n\n## Issue Transformer Tests (tests/issue_transformer_tests.rs)\n\n- transforms_gitlab_issue_to_normalized_schema\n- extracts_labels_from_issue_payload\n- handles_missing_optional_fields_gracefully\n- converts_iso_timestamps_to_ms_epoch\n- sets_last_seen_at_to_current_time\n\n## Discussion Transformer Tests (tests/discussion_transformer_tests.rs)\n\n- transforms_discussion_payload_to_normalized_schema\n- extracts_notes_array_from_discussion\n- sets_individual_note_flag_correctly\n- flags_system_notes_with_is_system_true\n- preserves_note_order_via_position_field\n- computes_first_note_at_and_last_note_at_correctly\n- computes_resolvable_and_resolved_status\n\n## Test Setup\n- Load from test fixtures\n- Use serde_json for deserialization\n- Compare against expected NormalizedX structs\n\nFiles: tests/issue_transformer_tests.rs, tests/discussion_transformer_tests.rs\nDone when: All transformer unit tests pass","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:59:04.165187Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:02.015847Z","deleted_at":"2026-01-25T17:02:02.015841Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2mz","title":"Epic: Gate A - Lexical MVP","description":"## Background\nGate A delivers the lexical search MVP — the foundation that works without sqlite-vec or Ollama. It introduces the document layer (documents, document_labels, document_paths), FTS5 indexing, search filters, and the search + stats + generate-docs CLI commands. Gate A is independently shippable — users get working search with FTS5 only.\n\n## Gate A Deliverables\n1. Document generation from issues/MRs/discussions with FTS5 indexing\n2. Lexical search + filters + snippets + lore stats\n\n## Bead Dependencies (execution order)\n1. **bd-3lc** — Rename GiError to LoreError (no deps, enables all subsequent work)\n2. **bd-hrs** — Migration 007 (blocked by bd-3lc)\n3. **bd-221** — Migration 008 FTS5 (blocked by bd-hrs)\n4. **bd-36p** — Document types + extractor module (blocked by bd-3lc)\n5. **bd-18t** — Truncation logic (blocked by bd-36p)\n6. **bd-247** — Issue extraction (blocked by bd-36p, bd-hrs)\n7. **bd-1yz** — MR extraction (blocked by bd-36p, bd-hrs)\n8. **bd-2fp** — Discussion extraction (blocked by bd-36p, bd-hrs, bd-18t)\n9. **bd-1u1** — Document regenerator (blocked by bd-36p, bd-38q, bd-hrs)\n10. **bd-1k1** — FTS5 search (blocked by bd-221)\n11. **bd-3q2** — Search filters (blocked by bd-36p)\n12. **bd-3lu** — Search CLI (blocked by bd-1k1, bd-3q2, bd-36p)\n13. **bd-3qs** — Generate-docs CLI (blocked by bd-1u1, bd-3lu)\n14. **bd-pr1** — Stats CLI (blocked by bd-hrs)\n15. **bd-2dk** — Project resolution (blocked by bd-3lc)\n\n## Acceptance Criteria\n- [ ] `lore search \"query\"` returns FTS5 results with snippets\n- [ ] `lore search --type issue --label bug \"query\"` filters correctly\n- [ ] `lore generate-docs` creates documents from all entities\n- [ ] `lore generate-docs --full` regenerates everything\n- [ ] `lore stats` shows document/FTS/queue counts\n- [ ] `lore stats --check` verifies FTS consistency\n- [ ] No sqlite-vec dependency in Gate A","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-30T15:25:09.721108Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:54:44.243610Z","closed_at":"2026-01-30T17:54:44.243562Z","close_reason":"All Gate A sub-beads complete. Lexical MVP delivered: document extraction (issue/MR/discussion), FTS5 indexing, search with filters/snippets/RRF, generate-docs CLI, stats CLI with integrity check/repair.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2mz","depends_on_id":"bd-3lu","type":"blocks","created_at":"2026-01-30T15:29:35.679499Z","created_by":"tayloreernisse"},{"issue_id":"bd-2mz","depends_on_id":"bd-3qs","type":"blocks","created_at":"2026-01-30T15:29:35.713718Z","created_by":"tayloreernisse"},{"issue_id":"bd-2mz","depends_on_id":"bd-pr1","type":"blocks","created_at":"2026-01-30T15:29:35.747904Z","created_by":"tayloreernisse"}]}
+{"id":"bd-2n4","title":"Implement trace query: file -> MR -> issue -> discussion chain","description":"## Background\nlore trace builds a chain from file to motivating decisions (Tier 1: API-only, no git blame). Combines file-history (bd-z94) with cross-reference data (bd-3ia) to answer \"Why was this code introduced?\"\n\n## Approach\nCreate src/core/trace.rs:\n\n```rust\npub struct TraceResult {\n    pub file_path: String,\n    pub rename_chain: Option<Vec<String>>,\n    pub trace_chains: Vec<TraceChain>,\n}\n\npub struct TraceChain {\n    pub mr: TraceMR,\n    pub issues: Vec<TraceIssue>,\n    pub file_discussions: Vec<TraceDiscussion>,\n}\n\npub struct TraceMR {\n    pub iid: i64,\n    pub title: String,\n    pub state: String,\n    pub author: String,\n    pub merged_at: Option<i64>,\n    pub created_at: i64,\n    pub web_url: String,\n    pub change_type: String,\n    pub discussion_count: usize,\n    pub file_discussion_count: usize,\n}\n\npub struct TraceIssue {\n    pub iid: i64,\n    pub title: String,\n    pub state: String,\n    pub web_url: String,\n    pub reference_type: String,    // \"closes\" | \"related\"\n    pub discussion_count: usize,\n}\n\npub struct TraceDiscussion {\n    pub note_id: i64,\n    pub author: String,\n    pub snippet: String,           // first ~200 chars\n    pub position_path: String,\n    pub position_line: Option<i64>,\n    pub created_at: i64,\n}\n\n/// Build trace chains for a file path.\npub fn build_trace(\n    conn: &Connection,\n    path: &str,\n    project_id: Option<i64>,\n    follow_renames: bool,\n) -> Result<TraceResult>\n```\n\nQuery flow (spec §5.4):\n1. Resolve rename chain (reuse file_history::resolve_rename_chain)\n2. Find MRs via mr_file_changes (same query as file-history)\n3. For each MR, find related issues via entity_references WHERE source=MR AND reference_type IN ('closes', 'related')\n4. For each issue, count discussions\n5. For each MR, fetch DiffNote discussions on the traced file:\n```sql\nSELECT n.id, n.author_username, n.body, n.position_new_path, n.position_new_line, n.created_at\nFROM notes n\nJOIN discussions d ON n.discussion_id = d.id\nWHERE d.merge_request_id = ?\n  AND (n.position_new_path IN (rename_chain_paths))\nORDER BY n.created_at;\n```\n6. Build trace chains, ordered by MR merge date (most recent first)\n\nRegister in src/core/mod.rs: `pub mod trace;`\n\n## Acceptance Criteria\n- [ ] Correct MR-to-issue linking via entity_references\n- [ ] DiffNote snippets for the traced file included\n- [ ] Multi-hop rename resolution applied\n- [ ] Chains ordered by merge date (most recent first)\n- [ ] Graceful when no MR data found\n- [ ] Issues show reference_type (closes vs related)\n\n## Files\n- src/core/trace.rs (new)\n- src/core/mod.rs (add `pub mod trace;`)\n\n## TDD Loop\nRED: tests/trace_tests.rs:\n- `test_trace_basic_chain` - MR→issue→discussions chain built correctly\n- `test_trace_with_file_discussions` - DiffNotes on the file included\n- `test_trace_no_data` - empty result, not error\n- `test_trace_multiple_mrs` - multiple MRs touching file, each with its own chain\n- `test_trace_rename_chain` - renamed file finds MRs at old path\n\nSetup: create_test_db with full migrations, seed issues + MRs + entity_references + notes with positions + mr_file_changes.\n\nGREEN: Implement build_trace\n\nVERIFY: `cargo test trace -- --nocapture`\n\n## Edge Cases\n- MR with no linked issues (no entity_references): chain has MR but empty issues list\n- DiffNote with position on OLD path (before rename): check position_old_path too\n- Very large repos: file may have 50+ MRs — respect caller's limit\n- Deleted issues: entity_reference target may point to deleted issue — skip gracefully","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:34:32.738743Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:49:51.190052Z","compaction_level":0,"original_size":0,"labels":["gate-5","phase-b","query"],"dependencies":[{"issue_id":"bd-2n4","depends_on_id":"bd-1ht","type":"parent-child","created_at":"2026-02-02T21:34:32.743943Z","created_by":"tayloreernisse"},{"issue_id":"bd-2n4","depends_on_id":"bd-3ia","type":"blocks","created_at":"2026-02-02T21:34:37.899870Z","created_by":"tayloreernisse"},{"issue_id":"bd-2n4","depends_on_id":"bd-z94","type":"blocks","created_at":"2026-02-02T21:34:37.854791Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2nb","title":"[CP1] Issue ingestion module","description":"Fetch and store issues with cursor-based incremental sync.\n\nImplement ingestIssues(options) → { fetched, upserted, labelsCreated }\n\nLogic:\n1. Get current cursor from sync_cursors\n2. Paginate through issues updated after cursor\n3. Apply local filtering for tuple cursor semantics\n4. For each issue:\n   - Store raw payload (compressed)\n   - Upsert issue record\n   - Extract and upsert labels\n   - Link issue to labels via junction\n5. Update cursor after each page commit\n\nFiles: src/ingestion/issues.ts\nTests: tests/integration/issue-ingestion.test.ts\nDone when: Issues, labels, issue_labels populated correctly with resumable cursor","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:19:50.701180Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.154318Z","deleted_at":"2026-01-25T15:21:35.154316Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2no","title":"Write integration tests","description":"## Background\nIntegration tests verify that modules work together with a real SQLite database. They test FTS search (stemming, empty results), embedding storage (sqlite-vec ops), hybrid search (combined retrieval), and sync orchestration (full pipeline). Each test creates a fresh in-memory DB with migrations applied.\n\n## Approach\nCreate integration test files in `tests/`:\n\n**1. tests/fts_search.rs:**\n- Create DB, apply migrations 001-008\n- Insert test documents via SQL\n- Verify FTS5 triggers fired (documents_fts has matching count)\n- Search with various queries: stemming, prefix, empty, special chars\n- Verify result ranking (BM25 ordering)\n- Verify snippet generation\n\n**2. tests/embedding.rs:**\n- Create DB, apply migrations 001-009 (requires sqlite-vec)\n- Insert test documents + embeddings with known vectors\n- Verify KNN search returns nearest neighbors\n- Verify chunk deduplication\n- Verify orphan cleanup trigger (delete document -> embeddings gone)\n\n**3. tests/hybrid_search.rs:**\n- Create DB, apply all migrations\n- Insert documents + embeddings\n- Test all three modes: lexical, semantic, hybrid\n- Verify RRF ranking produces expected order\n- Test graceful degradation (no embeddings -> FTS fallback)\n- Test adaptive recall with filters\n\n**4. tests/sync.rs:**\n- Test sync orchestration with mock/stub GitLab responses\n- Verify pipeline stages execute in order\n- Verify lock acquisition/release\n- Verify --no-embed and --no-docs flags\n\n**Test fixtures:**\n- Deterministic embedding vectors (no Ollama required): e.g., [1.0, 0.0, 0.0, ...] for doc1, [0.0, 1.0, 0.0, ...] for doc2\n- Known documents with predictable search results\n- Fixed timestamps for reproducibility\n\n## Acceptance Criteria\n- [ ] FTS search tests pass (stemming, prefix, empty, special chars)\n- [ ] Embedding tests pass (KNN, dedup, orphan cleanup)\n- [ ] Hybrid search tests pass (all 3 modes, graceful degradation)\n- [ ] Sync tests pass (pipeline orchestration)\n- [ ] All tests use in-memory DB (no file I/O)\n- [ ] No external dependencies (no Ollama, no GitLab) — use fixtures/stubs\n- [ ] `cargo test --test fts_search --test embedding --test hybrid_search --test sync` passes\n\n## Files\n- `tests/fts_search.rs` — new file\n- `tests/embedding.rs` — new file\n- `tests/hybrid_search.rs` — new file\n- `tests/sync.rs` — new file\n- `tests/fixtures/` — optional: test helper functions (shared DB setup)\n\n## TDD Loop\nThese ARE integration tests — they verify the combined behavior of multiple beads.\nVERIFY: `cargo test --test fts_search && cargo test --test embedding && cargo test --test hybrid_search && cargo test --test sync`\n\n## Edge Cases\n- sqlite-vec not available: embedding tests should skip gracefully (or require feature flag)\n- In-memory DB with WAL mode: may behave differently than file DB — test both if critical\n- Concurrent test execution: each test creates its own DB (no shared state)","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:27:21.751019Z","created_by":"tayloreernisse","updated_at":"2026-01-30T18:11:12.432092Z","closed_at":"2026-01-30T18:11:12.432036Z","close_reason":"Integration tests: 10 FTS search tests (stemming, empty, special chars, ordering, triggers, null title), 5 embedding tests (KNN, limit, dedup, orphan trigger, empty DB), 6 hybrid search tests (lexical mode, FTS-only, graceful degradation, RRF ranking, filters, mode variants). 310 total tests pass.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2no","depends_on_id":"bd-1x6","type":"blocks","created_at":"2026-01-30T15:29:35.607603Z","created_by":"tayloreernisse"},{"issue_id":"bd-2no","depends_on_id":"bd-3eu","type":"blocks","created_at":"2026-01-30T15:29:35.572825Z","created_by":"tayloreernisse"},{"issue_id":"bd-2no","depends_on_id":"bd-3lu","type":"blocks","created_at":"2026-01-30T15:29:35.499831Z","created_by":"tayloreernisse"},{"issue_id":"bd-2no","depends_on_id":"bd-am7","type":"blocks","created_at":"2026-01-30T15:29:35.535320Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2px","title":"[CP1] Epic: Issue Ingestion","description":"Ingest all issues, labels, and issue discussions from configured GitLab repositories with resumable cursor-based incremental sync. This establishes the core data ingestion pattern reused for MRs in CP2.\n\n## Success Criteria\n- gi ingest --type=issues fetches all issues (count matches GitLab UI)\n- Labels extracted from issue payloads (name-only)\n- Label linkage reflects current GitLab state (removed labels unlinked on re-sync)\n- Issue discussions fetched per-issue (dependent sync)\n- Cursor-based sync is resumable (re-running fetches 0 new items)\n- Discussion sync skips unchanged issues (per-issue watermark)\n- Sync tracking records all runs\n- Single-flight lock prevents concurrent runs\n\n## Internal Gates\n- Gate A: Issues only (cursor + upsert + raw payloads + list/count/show)\n- Gate B: Labels correct (stale-link removal verified)\n- Gate C: Dependent discussion sync (watermark prevents redundant refetch)\n- Gate D: Resumability proof (kill mid-run, rerun; bounded redo)\n\nReference: docs/prd/checkpoint-1.md","status":"tombstone","priority":1,"issue_type":"epic","created_at":"2026-01-25T15:42:13.167698Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.638609Z","deleted_at":"2026-01-25T17:02:01.638606Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"epic","compaction_level":0,"original_size":0}
 {"id":"bd-2sx","title":"Implement lore embed CLI command","description":"## Background\nThe embed CLI command is the user-facing wrapper for the embedding pipeline. It runs Ollama health checks, selects documents to embed (pending or failed), shows progress, and reports results. This is the standalone command for building embeddings outside of the sync orchestrator.\n\n## Approach\nCreate `src/cli/commands/embed.rs` per PRD Section 4.4.\n\n**IMPORTANT: The embed command is async.** The underlying `embed_documents()` function is `async fn` (uses `FuturesUnordered` for concurrent HTTP to Ollama). The CLI runner must use tokio runtime.\n\n**Core function (async):**\n```rust\npub async fn run_embed(\n    config: &Config,\n    retry_failed: bool,\n) -> Result<EmbedResult>\n```\n\n**Pipeline:**\n1. Create OllamaClient from config.embedding (base_url, model, timeout_secs)\n2. Run `client.health_check().await` — fail early with clear error if Ollama unavailable or model missing\n3. Determine selection: `EmbedSelection::RetryFailed` if --retry-failed, else `EmbedSelection::Pending`\n4. Call `embed_documents(conn, &client, selection, concurrency, progress_callback).await`\n   - `concurrency` param controls max in-flight HTTP requests to Ollama\n   - `progress_callback` drives indicatif progress bar\n5. Show progress bar (indicatif) during embedding\n6. Return EmbedResult with counts\n\n**CLI args:**\n```rust\n#[derive(Args)]\npub struct EmbedArgs {\n    #[arg(long)]\n    retry_failed: bool,\n}\n```\n\n**Output:**\n- Human: \"Embedded 42 documents (15 chunks), 2 errors, 5 skipped (unchanged)\"\n- JSON: `{\"ok\": true, \"data\": {\"embedded\": 42, \"chunks\": 15, \"errors\": 2, \"skipped\": 5}}`\n\n**Tokio integration note:**\nThe embed command runs async code. Either:\n- Use `#[tokio::main]` on main and propagate async through CLI dispatch\n- Or use `tokio::runtime::Runtime::new()` in the embed command handler\n\n## Acceptance Criteria\n- [ ] Command is async (embed_documents is async, health_check is async)\n- [ ] OllamaClient created from config.embedding settings\n- [ ] Health check runs first — clear error if Ollama down (exit code 14)\n- [ ] Clear error if model not found: \"Pull the model: ollama pull nomic-embed-text\" (exit code 15)\n- [ ] Embeds pending documents (no existing embeddings or stale content_hash)\n- [ ] --retry-failed re-attempts documents with last_error\n- [ ] Progress bar shows during embedding (indicatif)\n- [ ] embed_documents called with concurrency parameter\n- [ ] embed_documents called with progress_callback for progress bar\n- [ ] Human + JSON output\n- [ ] `cargo build` succeeds\n\n## Files\n- `src/cli/commands/embed.rs` — new file\n- `src/cli/commands/mod.rs` — add `pub mod embed;`\n- `src/cli/mod.rs` — add EmbedArgs, wire up embed subcommand\n- `src/main.rs` — add embed command handler (async dispatch)\n\n## TDD Loop\nRED: Integration test needing Ollama\nGREEN: Implement run_embed (async)\nVERIFY: `cargo build && cargo test embed`\n\n## Edge Cases\n- No documents in DB: \"No documents to embed\" (not error)\n- All documents already embedded and unchanged: \"0 documents to embed (all up to date)\"\n- Ollama goes down mid-embedding: pipeline records errors for remaining docs, returns partial result\n- --retry-failed with no failed docs: \"No failed documents to retry\"","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:34.126482Z","created_by":"tayloreernisse","updated_at":"2026-01-30T18:02:38.633115Z","closed_at":"2026-01-30T18:02:38.633055Z","close_reason":"Embed CLI command fully wired: EmbedArgs, Commands::Embed variant, handle_embed handler, clean build, all tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-2sx","depends_on_id":"bd-am7","type":"blocks","created_at":"2026-01-30T15:29:24.766104Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2ug","title":"[CP1] gi ingest --type=issues command","description":"CLI command to orchestrate issue ingestion.\n\n## Module\nsrc/cli/commands/ingest.rs\n\n## Clap Definition\n#[derive(Subcommand)]\npub enum Commands {\n    Ingest {\n        #[arg(long, value_parser = [\"issues\", \"merge_requests\"])]\n        r#type: String,\n        \n        #[arg(long)]\n        project: Option<String>,\n        \n        #[arg(long)]\n        force: bool,\n    },\n}\n\n## Implementation\n1. Acquire app lock with heartbeat (respect --force for stale lock)\n2. Create sync_run record (status='running')\n3. For each configured project (or filtered --project):\n   - Call orchestrator to ingest issues and discussions\n   - Show progress (spinner or progress bar)\n4. Update sync_run (status='succeeded', metrics_json with counts)\n5. Release lock\n\n## Output Format\nIngesting issues...\n\n  group/project-one: 1,234 issues fetched, 45 new labels\n\nFetching discussions (312 issues with updates)...\n\n  group/project-one: 312 issues → 1,234 discussions, 5,678 notes\n\nTotal: 1,234 issues, 1,234 discussions, 5,678 notes (excluding 1,234 system notes)\nSkipped discussion sync for 922 unchanged issues.\n\n## Error Handling\n- Lock acquisition failure: exit with DatabaseLockError message\n- Network errors: show GitLabNetworkError, exit non-zero\n- Rate limiting: respect backoff, show progress\n\nFiles: src/cli/commands/ingest.rs, src/cli/commands/mod.rs\nTests: tests/integration/sync_runs_tests.rs\nDone when: Full issue + discussion ingestion works end-to-end","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T16:57:58.552504Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.875613Z","deleted_at":"2026-01-25T17:02:01.875607Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2um","title":"[CP1] Epic: Issue Ingestion","description":"Ingest all issues, labels, and issue discussions from configured GitLab repositories with resumable cursor-based incremental sync. This checkpoint establishes the core data ingestion pattern that will be reused for MRs in Checkpoint 2.\n\n## Success Criteria\n- gi ingest --type=issues fetches all issues (count matches GitLab UI)\n- Labels extracted from issue payloads (name-only)\n- Label linkage reflects current GitLab state (removed labels unlinked on re-sync)\n- Issue discussions fetched per-issue (dependent sync)\n- Cursor-based sync is resumable (re-running fetches 0 new items)\n- Discussion sync skips unchanged issues (per-issue watermark)\n- Sync tracking records all runs (sync_runs table)\n- Single-flight lock prevents concurrent runs\n\n## Internal Gates\n- **Gate A**: Issues only - cursor + upsert + raw payloads + list/count/show working\n- **Gate B**: Labels correct - stale-link removal verified; label count matches GitLab\n- **Gate C**: Dependent discussion sync - watermark prevents redundant refetch; concurrency bounded\n- **Gate D**: Resumability proof - kill mid-run, rerun; bounded redo and no redundant discussion refetch\n\n## Reference\ndocs/prd/checkpoint-1.md","status":"closed","priority":1,"issue_type":"epic","created_at":"2026-01-25T17:02:38.075224Z","created_by":"tayloreernisse","updated_at":"2026-01-25T23:27:15.347364Z","closed_at":"2026-01-25T23:27:15.347317Z","close_reason":"CP1 Issue Ingestion complete: all sub-tasks done, 71 tests pass, CLI commands working","compaction_level":0,"original_size":0}
+{"id":"bd-2yo","title":"Fetch MR diffs API and populate mr_file_changes","description":"## Background\nGET /projects/:id/merge_requests/:iid/diffs returns file change metadata. We extract file paths and change types but NOT diff content. Uses the generic dependent fetch queue (job_type = 'mr_diffs').\n\n## Approach\n\n**1. Add API endpoint (src/gitlab/client.rs):**\n```rust\npub async fn fetch_mr_diffs(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabDiffFile>>\n```\n\nNew type in src/gitlab/types.rs:\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabDiffFile {\n    pub old_path: String,\n    pub new_path: String,\n    pub new_file: bool,\n    pub renamed_file: bool,\n    pub deleted_file: bool,\n    // diff content fields exist but we ignore them\n}\n```\n\nURL: `GET /api/v4/projects/{project_id}/merge_requests/{iid}/diffs?per_page=100`\n\n**2. Enqueue jobs during MR ingestion:**\nIn orchestrator.rs, after MR upsert (when fetch_mr_file_changes is true):\n```rust\nif config.sync.fetch_mr_file_changes {\n    enqueue_job(conn, project_id, \"merge_request\", iid, local_id, \"mr_diffs\", None)?;\n}\n```\n\n**3. Process jobs in drain step:**\nHandle \"mr_diffs\" job_type:\n```rust\nlet diffs = client.fetch_mr_diffs(gitlab_project_id, job.entity_iid).await?;\n// DELETE existing rows for this MR (diffs can change on rebase)\nconn.execute(\"DELETE FROM mr_file_changes WHERE merge_request_id = ?\", [job.entity_local_id])?;\n// Insert new rows\nfor diff in &diffs {\n    let change_type = if diff.new_file { \"added\" }\n        else if diff.renamed_file { \"renamed\" }\n        else if diff.deleted_file { \"deleted\" }\n        else { \"modified\" };\n    conn.execute(\n        \"INSERT INTO mr_file_changes (merge_request_id, project_id, old_path, new_path, change_type) VALUES (?, ?, ?, ?, ?)\",\n        params![job.entity_local_id, project_id, \n                if diff.renamed_file { Some(&diff.old_path) } else { None },\n                &diff.new_path, change_type],\n    )?;\n}\n```\n\n**4. Also capture commit SHAs:**\nDuring MR ingestion (orchestrator.rs), update merge_requests with merge_commit_sha and squash_commit_sha from the GitLab API response. These fields need to be added to GitLabMergeRequest type and transformer.\n\nAdd to src/gitlab/types.rs GitLabMergeRequest:\n```rust\npub merge_commit_sha: Option<String>,\npub squash_commit_sha: Option<String>,\n```\n\nUpdate MR transformer to pass these through, and UPDATE merge_requests SET merge_commit_sha = ?, squash_commit_sha = ? during upsert.\n\n## Acceptance Criteria\n- [ ] fetch_mr_diffs returns file metadata (no diff content)\n- [ ] Change types correctly derived: new_file→added, renamed_file→renamed, deleted_file→deleted, else→modified\n- [ ] Re-sync DELETEs + re-inserts (handles rebased MRs)\n- [ ] old_path only populated for renamed files\n- [ ] merge_commit_sha and squash_commit_sha captured in merge_requests table\n- [ ] Jobs only enqueued when fetch_mr_file_changes is true\n\n## Files\n- src/gitlab/client.rs (add fetch_mr_diffs)\n- src/gitlab/types.rs (add GitLabDiffFile, add fields to GitLabMergeRequest)\n- src/gitlab/transformers/merge_request.rs (pass through commit SHAs)\n- src/ingestion/orchestrator.rs (enqueue mr_diffs jobs, update commit SHAs)\n- src/core/drain.rs or sync.rs (handle mr_diffs in drain dispatcher)\n\n## TDD Loop\nRED: tests/file_changes_tests.rs:\n- `test_derive_change_type_added` - new_file=true → \"added\"\n- `test_derive_change_type_renamed` - renamed_file=true → \"renamed\", old_path populated\n- `test_derive_change_type_deleted` - deleted_file=true → \"deleted\"\n- `test_derive_change_type_modified` - all false → \"modified\"\n- `test_resync_deletes_and_reinserts` - insert, then re-process with different files, verify old rows gone\n\ntests/gitlab_types_tests.rs:\n- `test_deserialize_diff_file` - verify GitLabDiffFile deserialization\n- `test_deserialize_mr_with_commit_shas` - verify new fields on GitLabMergeRequest\n\nGREEN: Implement API endpoint, change type derivation, drain handler, commit SHA capture\n\nVERIFY: `cargo test file_changes -- --nocapture && cargo test gitlab_types -- --nocapture`\n\n## Edge Cases\n- MR with 1000+ files (monorepo): pagination essential on diffs endpoint\n- old_path for non-renames: GitLab still returns old_path (same as new_path) — only store when renamed_file=true\n- Draft MRs: diffs may change frequently — DELETE+INSERT handles this\n- MR with no diffs (empty MR): returns empty array, no rows inserted, job still completed\n- merge_commit_sha is NULL until MR is merged — don't error on NULL","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:34:08.939514Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:48:37.319521Z","compaction_level":0,"original_size":0,"labels":["api","gate-4","phase-b"],"dependencies":[{"issue_id":"bd-2yo","depends_on_id":"bd-14q","type":"parent-child","created_at":"2026-02-02T21:34:08.941359Z","created_by":"tayloreernisse"},{"issue_id":"bd-2yo","depends_on_id":"bd-1oo","type":"blocks","created_at":"2026-02-02T21:34:16.555239Z","created_by":"tayloreernisse"},{"issue_id":"bd-2yo","depends_on_id":"bd-jec","type":"blocks","created_at":"2026-02-02T21:34:16.656402Z","created_by":"tayloreernisse"},{"issue_id":"bd-2yo","depends_on_id":"bd-tir","type":"blocks","created_at":"2026-02-02T21:34:16.605198Z","created_by":"tayloreernisse"}]}
 {"id":"bd-2yq","title":"[CP1] Issue transformer with label extraction","description":"Transform GitLab issue payloads to normalized database schema.\n\nFunctions to implement:\n- transformIssue(gitlabIssue, localProjectId) → NormalizedIssue\n- extractLabels(gitlabIssue, localProjectId) → Label[]\n\nTransformation rules:\n- Convert ISO timestamps to ms epoch using isoToMs()\n- Set last_seen_at to nowMs()\n- Handle labels vs labels_details (prefer details when available)\n- Handle missing optional fields gracefully\n\nFiles: src/gitlab/transformers/issue.ts\nTests: tests/unit/issue-transformer.test.ts\nDone when: Unit tests pass for payload transformation and label extraction","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:19:09.660448Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.152259Z","deleted_at":"2026-01-25T15:21:35.152254Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-2ys","title":"[CP1] Cargo.toml updates - async-stream and futures","description":"## Background\n\nThe GitLab client pagination methods require async streaming capabilities. The `async-stream` crate provides the `stream!` macro for creating async iterators, and `futures` provides `StreamExt` for consuming them with `.next()` and other combinators.\n\n## Approach\n\nAdd these dependencies to Cargo.toml:\n\n```toml\n[dependencies]\nasync-stream = \"0.3\"\nfutures = { version = \"0.3\", default-features = false, features = [\"alloc\"] }\n```\n\nUse minimal features on `futures` to avoid pulling unnecessary code.\n\n## Acceptance Criteria\n\n- [ ] `async-stream = \"0.3\"` is in Cargo.toml [dependencies]\n- [ ] `futures` with `alloc` feature is in Cargo.toml [dependencies]\n- [ ] `cargo check` succeeds after adding dependencies\n\n## Files\n\n- Cargo.toml (edit)\n\n## TDD Loop\n\nRED: Not applicable (dependency addition)\nGREEN: Add lines to Cargo.toml\nVERIFY: `cargo check`\n\n## Edge Cases\n\n- If `futures` is already present, merge features rather than duplicate\n- Use exact version pins for reproducibility","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.104664Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:25:10.274787Z","closed_at":"2026-01-25T22:25:10.274727Z","close_reason":"Added async-stream 0.3 and futures 0.3 (alloc feature) to Cargo.toml, cargo check passes","compaction_level":0,"original_size":0}
+{"id":"bd-2zl","title":"Epic: Gate 1 - Resource Events Ingestion","description":"Ingest structured event data from GitLab Resource Events APIs (state, label, milestone) into local event tables. Includes migration 011 (renumbered from spec's 010 since 010_chunk_config.sql already exists), config extension, API client endpoints, generic dependent fetch queue, and ingestion pipeline integration.\n\nChildren: bd-hu3 (migration 011), bd-2e8 (config), bd-2fm (types), bd-sqw (API), bd-1uc (DB upserts), bd-tir (queue), bd-1ep (pipeline wiring), bd-3sh (count events), bd-1m8 (stats check)","status":"open","priority":1,"issue_type":"feature","created_at":"2026-02-02T21:30:49.136036Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:40.551205Z","compaction_level":0,"original_size":0,"labels":["epic","gate-1","phase-b"]}
 {"id":"bd-2zr","title":"[CP1] GitLab client pagination methods","description":"Add async stream methods for paginated GitLab API calls.\n\n## Methods to Add to GitLabClient\n\n### paginate_issues(gitlab_project_id, updated_after, cursor_rewind_seconds) -> Stream<GitLabIssue>\n- Use async_stream::try_stream! macro\n- Query params: scope=all, state=all, order_by=updated_at, sort=asc, per_page=100\n- If updated_after provided, apply cursor_rewind_seconds (subtract from timestamp)\n- Clamp to 0 to avoid underflow: (ts - rewind_ms).max(0)\n- Follow x-next-page header until empty/absent\n- Fall back to empty-page detection if headers missing\n\n### paginate_issue_discussions(gitlab_project_id, issue_iid) -> Stream<GitLabDiscussion>\n- Paginate through discussions for single issue\n- per_page=100\n- Follow x-next-page header\n\n### request_with_headers<T>(path, params) -> Result<(T, HeaderMap)>\n- Acquire rate limiter\n- Make request with PRIVATE-TOKEN header\n- Return both deserialized data and response headers\n\n## Dependencies\n- async-stream = \"0.3\" (for try_stream! macro)\n- futures = \"0.3\" (for Stream trait and StreamExt)\n\nFiles: src/gitlab/client.rs\nTests: tests/pagination_tests.rs\nDone when: Pagination handles multiple pages and respects cursors, tests pass","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T16:57:13.045971Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.784887Z","deleted_at":"2026-01-25T17:02:01.784883Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-31b","title":"[CP1] Discussion ingestion module","description":"Fetch and store discussions/notes for each issue.\n\nImplement ingestIssueDiscussions(options) → { discussionsFetched, discussionsUpserted, notesUpserted, systemNotesCount }\n\nLogic:\n1. Paginate through all discussions for given issue\n2. For each discussion:\n   - Store raw payload (compressed)\n   - Upsert discussion record with correct issue FK\n   - Transform and upsert all notes\n   - Store raw payload per note\n   - Track system notes count\n\nFiles: src/ingestion/discussions.ts\nTests: tests/integration/issue-discussion-ingestion.test.ts\nDone when: Discussions and notes populated with correct FKs and is_system flags","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:19:57.131442Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.156574Z","deleted_at":"2026-01-25T15:21:35.156571Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-31i","title":"Epic: CP2 Gate B - Labels + Assignees + Reviewers","description":"## Background\nGate B validates junction tables for labels, assignees, and reviewers. Ensures relationships are tracked correctly and stale links are removed on resync. This is critical for filtering (`--reviewer=alice`) and display.\n\n## Acceptance Criteria (Pass/Fail)\n- [ ] `mr_labels` table has rows for MRs with labels\n- [ ] Label count per MR matches GitLab UI (spot check 3 MRs)\n- [ ] `mr_assignees` table has rows for MRs with assignees\n- [ ] Assignee usernames match GitLab UI (spot check 3 MRs)\n- [ ] `mr_reviewers` table has rows for MRs with reviewers\n- [ ] Reviewer usernames match GitLab UI (spot check 3 MRs)\n- [ ] Remove label in GitLab -> resync -> link removed from mr_labels\n- [ ] Add reviewer in GitLab -> resync -> link added to mr_reviewers\n- [ ] `gi list mrs --label=bugfix` filters correctly\n- [ ] `gi list mrs --reviewer=alice` filters correctly\n\n## Validation Script\n```bash\n#!/bin/bash\nset -e\n\nDB_PATH=\"${XDG_DATA_HOME:-$HOME/.local/share}/gitlab-inbox/db.sqlite3\"\n\necho \"=== Gate B: Labels + Assignees + Reviewers ===\"\n\n# 1. Check label linkage exists\necho \"Step 1: Check label linkage...\"\nLABEL_LINKS=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM mr_labels;\")\necho \"  Total label links: $LABEL_LINKS\"\n\n# 2. Show sample label linkage\necho \"Step 2: Sample label linkage...\"\nsqlite3 \"$DB_PATH\" \"\n  SELECT m.iid, GROUP_CONCAT(l.name, ', ') as labels\n  FROM merge_requests m\n  JOIN mr_labels ml ON ml.merge_request_id = m.id\n  JOIN labels l ON l.id = ml.label_id\n  GROUP BY m.id\n  LIMIT 5;\n\"\n\n# 3. Check assignee linkage\necho \"Step 3: Check assignee linkage...\"\nASSIGNEE_LINKS=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM mr_assignees;\")\necho \"  Total assignee links: $ASSIGNEE_LINKS\"\n\n# 4. Show sample assignee linkage\necho \"Step 4: Sample assignee linkage...\"\nsqlite3 \"$DB_PATH\" \"\n  SELECT m.iid, GROUP_CONCAT(a.username, ', ') as assignees\n  FROM merge_requests m\n  JOIN mr_assignees a ON a.merge_request_id = m.id\n  GROUP BY m.id\n  LIMIT 5;\n\"\n\n# 5. Check reviewer linkage\necho \"Step 5: Check reviewer linkage...\"\nREVIEWER_LINKS=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM mr_reviewers;\")\necho \"  Total reviewer links: $REVIEWER_LINKS\"\n\n# 6. Show sample reviewer linkage\necho \"Step 6: Sample reviewer linkage...\"\nsqlite3 \"$DB_PATH\" \"\n  SELECT m.iid, GROUP_CONCAT(r.username, ', ') as reviewers\n  FROM merge_requests m\n  JOIN mr_reviewers r ON r.merge_request_id = m.id\n  GROUP BY m.id\n  LIMIT 5;\n\"\n\n# 7. Test filter commands\necho \"Step 7: Test filter commands...\"\n# Get a label that exists\nLABEL=$(sqlite3 \"$DB_PATH\" \"SELECT name FROM labels LIMIT 1;\")\nif [ -n \"$LABEL\" ]; then\n  echo \"  Testing --label=$LABEL\"\n  gi list mrs --label=\"$LABEL\" --limit=3\nfi\n\n# Get a reviewer that exists\nREVIEWER=$(sqlite3 \"$DB_PATH\" \"SELECT username FROM mr_reviewers LIMIT 1;\")\nif [ -n \"$REVIEWER\" ]; then\n  echo \"  Testing --reviewer=$REVIEWER\"\n  gi list mrs --reviewer=\"$REVIEWER\" --limit=3\nfi\n\necho \"\"\necho \"=== Gate B: PASSED ===\"\n```\n\n## Stale Link Removal Test (Manual)\n```bash\n# 1. Pick an MR with labels in GitLab UI\nMR_IID=123\n\n# 2. Note current label count\nsqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM mr_labels ml\n  JOIN merge_requests m ON m.id = ml.merge_request_id\n  WHERE m.iid = $MR_IID;\n\"\n# Example: 3 labels\n\n# 3. Remove a label in GitLab UI (manually)\n\n# 4. Resync\ngi ingest --type=merge_requests\n\n# 5. Verify label removed\nsqlite3 \"$DB_PATH\" \"\n  SELECT COUNT(*) FROM mr_labels ml\n  JOIN merge_requests m ON m.id = ml.merge_request_id\n  WHERE m.iid = $MR_IID;\n\"\n# Should be: 2 labels (one less)\n```\n\n## Test Commands (Quick Verification)\n```bash\n# Check counts:\nsqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"\n  SELECT \n    (SELECT COUNT(*) FROM mr_labels) as label_links,\n    (SELECT COUNT(*) FROM mr_assignees) as assignee_links,\n    (SELECT COUNT(*) FROM mr_reviewers) as reviewer_links;\n\"\n\n# Test filtering:\ngi list mrs --label=enhancement --limit=5\ngi list mrs --reviewer=alice --limit=5\ngi list mrs --assignee=bob --limit=5\n```\n\n## Dependencies\nThis gate requires:\n- bd-ser (MR ingestion with label/assignee/reviewer linking via clear-and-relink pattern)\n- Gate A must pass first\n\n## Edge Cases\n- MRs with no labels/assignees/reviewers: junction tables should have no rows for that MR\n- Labels shared across issues and MRs: labels table is shared, only junction differs\n- Usernames are case-sensitive: `Alice` != `alice`","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-26T22:06:01.292318Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:48:21.059422Z","closed_at":"2026-01-27T00:48:21.059378Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-31i","depends_on_id":"bd-ser","type":"blocks","created_at":"2026-01-26T22:08:55.684769Z","created_by":"tayloreernisse"}]}
 {"id":"bd-31m","title":"[CP1] Test fixtures for mocked GitLab responses","description":"Create mock response files for integration tests.\n\nFixtures to create:\n- gitlab-issue.json (single issue with labels)\n- gitlab-issues-page.json (paginated list)\n- gitlab-discussion.json (single discussion with notes)\n- gitlab-discussions-page.json (paginated list)\n\nInclude edge cases:\n- Issue with labels_details\n- Issue with no labels\n- Discussion with individual_note=true\n- System notes with system=true\n\nFiles: tests/fixtures/mock-responses/gitlab-issue*.json, gitlab-discussion*.json\nDone when: MSW handlers can use fixtures for deterministic tests","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T15:20:43.781288Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.155480Z","deleted_at":"2026-01-25T15:21:35.155478Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-327","title":"[CP0] Project scaffold","description":"## Background\n\nThis is the foundational scaffold for the GitLab Inbox CLI tool. Every subsequent bead depends on having the correct project structure, TypeScript configuration, and tooling in place. The configuration choices here (ESM modules, strict TypeScript, Vitest for testing) set constraints for all future code.\n\n## Approach\n\nCreate a Node.js 20+ ESM project with TypeScript strict mode. Use flat ESLint config (v9+) with TypeScript plugin. Configure Vitest with coverage. Create the directory structure matching the PRD exactly.\n\n**package.json essentials:**\n- `\"type\": \"module\"` for ESM\n- `\"bin\": { \"gi\": \"./dist/cli/index.js\" }` for CLI entry point\n- Runtime deps: better-sqlite3, sqlite-vec, commander, zod, pino, pino-pretty, ora, chalk, cli-table3, inquirer\n- Dev deps: typescript, @types/better-sqlite3, @types/node, vitest, msw, eslint, @typescript-eslint/*\n\n**tsconfig.json:**\n- `target: ES2022`, `module: Node16`, `moduleResolution: Node16`\n- `strict: true`, `noImplicitAny: true`, `strictNullChecks: true`\n- `outDir: ./dist`, `rootDir: ./src`\n\n**vitest.config.ts:**\n- Exclude `tests/live/**` unless `GITLAB_LIVE_TESTS=1`\n- Coverage with v8 provider\n\n## Acceptance Criteria\n\n- [ ] `npm install` completes without errors\n- [ ] `npm run build` compiles TypeScript to dist/\n- [ ] `npm run test` runs vitest (0 tests is fine at this stage)\n- [ ] `npm run lint` runs ESLint with no config errors\n- [ ] All directories exist: src/cli/commands/, src/core/, src/gitlab/, src/types/, tests/unit/, tests/integration/, tests/live/, tests/fixtures/mock-responses/, migrations/\n\n## Files\n\nCREATE:\n- package.json\n- tsconfig.json\n- vitest.config.ts\n- eslint.config.js\n- .gitignore\n- src/cli/index.ts (empty placeholder with shebang)\n- src/cli/commands/.gitkeep\n- src/core/.gitkeep\n- src/gitlab/.gitkeep\n- src/types/index.ts (empty)\n- tests/unit/.gitkeep\n- tests/integration/.gitkeep\n- tests/live/.gitkeep\n- tests/fixtures/mock-responses/.gitkeep\n- migrations/.gitkeep\n\n## TDD Loop\n\nN/A - scaffold only. Verify with:\n\n```bash\nnpm install\nnpm run build\nnpm run lint\nnpm run test\n```\n\n## Edge Cases\n\n- Node.js version < 20 will fail on ESM features - add `engines` field\n- better-sqlite3 requires native compilation - may need python/build-essential\n- sqlite-vec installation can fail on some platforms - document fallback","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:47.955044Z","created_by":"tayloreernisse","updated_at":"2026-01-25T02:51:25.347932Z","closed_at":"2026-01-25T02:51:25.347799Z","compaction_level":0,"original_size":0}
+{"id":"bd-32q","title":"Implement timeline seed phase: FTS5 keyword search to entity IDs","description":"## Background\nStep 1 of timeline query: find seed entities matching the keyword. Uses the existing FTS5 infrastructure from CP3 (documents_fts table). Documents map back to source entities (issues, MRs) via documents.source_type and source_id.\n\n## Approach\nAdd to src/core/timeline.rs:\n\n```rust\n/// Seed phase: FTS5 keyword search to find matching entities.\n/// Returns seed entities (issues, MRs) and top evidence notes.\npub fn seed_from_fts(\n    conn: &Connection,\n    query: &str,\n    project_id: Option<i64>,    // from -p flag\n    since: Option<i64>,          // ms epoch timestamp filter\n    limit: usize,                // -n flag\n) -> Result<SeedResult>\n\npub struct SeedResult {\n    pub entities: Vec<SeedEntity>,\n    pub evidence_notes: Vec<EvidenceNote>,\n}\n\npub struct SeedEntity {\n    pub entity_type: String,     // \"issue\" | \"merge_request\"\n    pub local_id: i64,           // DB id\n    pub iid: i64,\n    pub project_id: i64,\n    pub project_path: String,\n}\n\npub struct EvidenceNote {\n    pub note_id: i64,\n    pub discussion_id: Option<i64>,\n    pub entity_type: String,     // parent entity type\n    pub entity_iid: i64,\n    pub snippet: String,         // first ~200 chars\n    pub author: String,\n    pub created_at: i64,\n    pub url: Option<String>,\n    pub fts_rank: f64,\n}\n```\n\nSQL approach:\n1. FTS5 query on documents_fts:\n```sql\nSELECT d.id, d.source_type, d.source_id, d.project_id, d.title,\n       rank AS fts_rank\nFROM documents_fts\nJOIN documents d ON documents_fts.rowid = d.id\nWHERE documents_fts MATCH ?1\nORDER BY rank\nLIMIT ?2;\n```\n\n2. Map document results to source entities:\n   - source_type 'issue' → get issue by id\n   - source_type 'merge_request' → get MR by id\n   - source_type 'note' / 'discussion' → get parent entity\n\n3. Collect top 10 note matches as evidence candidates:\n   - Filter documents where source relates to a note/discussion\n   - Get note body, truncate to ~200 chars\n   - Record note_id, discussion_id, parent entity info\n\n4. Deduplicate entities (if a note match and its parent entity both appear)\n\n5. Apply filters: --since (filter by entity created_at), -p (project scope)\n\n## Acceptance Criteria\n- [ ] FTS5 search returns matching documents\n- [ ] Documents correctly mapped to source entities (issues, MRs)\n- [ ] Note matches produce evidence note entries (top 10 by FTS rank)\n- [ ] Deduplication: note match + parent entity don't double-count\n- [ ] --since filter applied to entity timestamps\n- [ ] -p filter scopes to project\n- [ ] Empty query returns error (not all entities)\n\n## Files\n- src/core/timeline.rs (add seed_from_fts function)\n\n## TDD Loop\nRED: tests/timeline_seed_tests.rs:\n- `test_seed_finds_matching_issue` - FTS match on issue title\n- `test_seed_finds_matching_note` - FTS match on note body, maps to parent entity\n- `test_seed_deduplicates_entities` - note match and issue match for same entity = 1 seed\n- `test_seed_collects_evidence_notes` - top notes returned with snippets\n- `test_seed_applies_since_filter` - old entities excluded\n- `test_seed_applies_project_filter` - wrong project excluded\n\nSetup: create_test_db with full migrations, seed documents + FTS index.\n\nGREEN: Implement seed_from_fts\n\nVERIFY: `cargo test timeline_seed -- --nocapture`\n\n## Edge Cases\n- FTS5 MATCH syntax: user may pass natural language — FTS5 handles this but special chars may need escaping\n- Documents table may have stale data (entity deleted but document remains) — skip missing entities\n- Note-type documents: need to map from note → discussion → parent entity. The documents table has source_type and source_id but the chain may be complex\n- Empty result set is valid (no matches) — return empty SeedResult, not error","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:08.615908Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:45:48.702921Z","compaction_level":0,"original_size":0,"labels":["gate-3","phase-b","query"],"dependencies":[{"issue_id":"bd-32q","depends_on_id":"bd-20e","type":"blocks","created_at":"2026-02-02T21:33:37.368005Z","created_by":"tayloreernisse"},{"issue_id":"bd-32q","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:08.617483Z","created_by":"tayloreernisse"}]}
 {"id":"bd-335","title":"Implement Ollama API client","description":"## Background\nThe Ollama API client provides the HTTP interface to the local Ollama embedding server. It handles health checks (is Ollama running? does the model exist?), batch embedding requests (up to 32 texts per call), and error translation to LoreError variants. This is the lowest-level embedding component — the pipeline (bd-am7) builds on top of it.\n\n## Approach\nCreate \\`src/embedding/ollama.rs\\` per PRD Section 4.2. **Uses async reqwest (not blocking).**\n\n```rust\nuse reqwest::Client;  // NOTE: async Client, not reqwest::blocking\nuse serde::{Deserialize, Serialize};\nuse crate::core::error::{LoreError, Result};\n\npub struct OllamaConfig {\n    pub base_url: String,      // default \\\"http://localhost:11434\\\"\n    pub model: String,         // default \\\"nomic-embed-text\\\"\n    pub timeout_secs: u64,     // default 60\n}\n\nimpl Default for OllamaConfig { /* PRD defaults */ }\n\npub struct OllamaClient {\n    client: Client,      // async reqwest::Client\n    config: OllamaConfig,\n}\n\n#[derive(Serialize)]\nstruct EmbedRequest { model: String, input: Vec<String> }\n\n#[derive(Deserialize)]\nstruct EmbedResponse { model: String, embeddings: Vec<Vec<f32>> }\n\n#[derive(Deserialize)]\nstruct TagsResponse { models: Vec<ModelInfo> }\n\n#[derive(Deserialize)]\nstruct ModelInfo { name: String }\n\nimpl OllamaClient {\n    pub fn new(config: OllamaConfig) -> Self;\n\n    /// Async health check: GET /api/tags\n    /// Model matched via starts_with (\\\"nomic-embed-text\\\" matches \\\"nomic-embed-text:latest\\\")\n    pub async fn health_check(&self) -> Result<()>;\n\n    /// Async batch embedding: POST /api/embed\n    /// Input: Vec<String> of texts, Response: Vec<Vec<f32>> of 768-dim embeddings\n    pub async fn embed_batch(&self, texts: Vec<String>) -> Result<Vec<Vec<f32>>>;\n}\n\n/// Quick health check without full client (async).\npub async fn check_ollama_health(base_url: &str) -> bool;\n```\n\n**Error mapping (per PRD):**\n- Connection refused/timeout -> LoreError::OllamaUnavailable { base_url, source: Some(e) }\n- Model not in /api/tags -> LoreError::OllamaModelNotFound { model }\n- Non-200 from /api/embed -> LoreError::EmbeddingFailed { document_id: 0, reason: format!(\\\"HTTP {}: {}\\\", status, body) }\n\n**Key PRD detail:** Model matching uses \\`starts_with\\` (not exact match) so \\\"nomic-embed-text\\\" matches \\\"nomic-embed-text:latest\\\".\n\n## Acceptance Criteria\n- [ ] Uses async reqwest::Client (not blocking)\n- [ ] health_check() is async, detects server availability and model presence\n- [ ] Model matched via starts_with (handles \\\":latest\\\" suffix)\n- [ ] embed_batch() is async, sends POST /api/embed\n- [ ] Batch size up to 32 texts\n- [ ] Returns Vec<Vec<f32>> with 768 dimensions each\n- [ ] OllamaUnavailable error includes base_url and source error\n- [ ] OllamaModelNotFound error includes model name\n- [ ] Non-200 response mapped to EmbeddingFailed with status + body\n- [ ] Timeout: 60 seconds default (configurable via OllamaConfig)\n- [ ] \\`cargo build\\` succeeds\n\n## Files\n- \\`src/embedding/ollama.rs\\` — new file\n- \\`src/embedding/mod.rs\\` — add \\`pub mod ollama;\\` and re-exports\n\n## TDD Loop\nRED: Tests (unit tests with mock, integration needs Ollama):\n- \\`test_config_defaults\\` — verify default base_url, model, timeout\n- \\`test_health_check_model_starts_with\\` — \\\"nomic-embed-text\\\" matches \\\"nomic-embed-text:latest\\\"\n- \\`test_embed_batch_parse\\` — mock response parsed correctly\n- \\`test_connection_error_maps_to_ollama_unavailable\\`\nGREEN: Implement OllamaClient\nVERIFY: \\`cargo test ollama\\`\n\n## Edge Cases\n- Ollama returns model name with version tag (\\\"nomic-embed-text:latest\\\"): starts_with handles this\n- Empty texts array: send empty batch, Ollama returns empty embeddings\n- Ollama returns wrong number of embeddings (2 texts, 1 embedding): caller (pipeline) validates\n- Non-JSON response: reqwest deserialization error -> wrap appropriately","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:34.025099Z","created_by":"tayloreernisse","updated_at":"2026-01-30T16:58:17.546852Z","closed_at":"2026-01-30T16:58:17.546794Z","close_reason":"Completed: OllamaClient with async health_check (starts_with model matching), embed_batch, error mapping to LoreError variants, check_ollama_health helper, 4 tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-335","depends_on_id":"bd-ljf","type":"blocks","created_at":"2026-01-30T15:29:24.627951Z","created_by":"tayloreernisse"}]}
 {"id":"bd-34o","title":"Implement MR transformer","description":"## Background\nTransforms GitLab MR API responses into normalized schema for database storage. Handles deprecated field fallbacks and extracts metadata (labels, assignees, reviewers).\n\n## Approach\nCreate new transformer module following existing issue transformer pattern:\n- `NormalizedMergeRequest` - Database-ready struct\n- `MergeRequestWithMetadata` - MR + extracted labels/assignees/reviewers\n- `transform_merge_request()` - Main transformation function\n- `extract_labels()` - Label extraction helper\n\n## Files\n- `src/gitlab/transformers/merge_request.rs` - New transformer module\n- `src/gitlab/transformers/mod.rs` - Export new module\n- `tests/mr_transformer_tests.rs` - Unit tests\n\n## Acceptance Criteria\n- [ ] `NormalizedMergeRequest` struct exists with all DB columns\n- [ ] `MergeRequestWithMetadata` contains MR + label_names + assignee_usernames + reviewer_usernames\n- [ ] `transform_merge_request()` returns `Result<MergeRequestWithMetadata, String>`\n- [ ] `draft` computed as `gitlab_mr.draft || gitlab_mr.work_in_progress`\n- [ ] `detailed_merge_status` prefers `detailed_merge_status` over `merge_status_legacy`\n- [ ] `merge_user_username` prefers `merge_user` over `merged_by`\n- [ ] `head_sha` extracted from `sha` field\n- [ ] `references_short` and `references_full` extracted from `references` Option\n- [ ] Timestamps parsed with `iso_to_ms()`, errors returned (not zeroed)\n- [ ] `last_seen_at` set to `now_ms()`\n- [ ] `cargo test mr_transformer` passes\n\n## TDD Loop\nRED: `cargo test mr_transformer` -> module not found\nGREEN: Add transformer with all fields\nVERIFY: `cargo test mr_transformer`\n\n## Struct Definitions\n```rust\n#[derive(Debug, Clone)]\npub struct NormalizedMergeRequest {\n    pub gitlab_id: i64,\n    pub project_id: i64,\n    pub iid: i64,\n    pub title: String,\n    pub description: Option<String>,\n    pub state: String,\n    pub draft: bool,\n    pub author_username: String,\n    pub source_branch: String,\n    pub target_branch: String,\n    pub head_sha: Option<String>,\n    pub references_short: Option<String>,\n    pub references_full: Option<String>,\n    pub detailed_merge_status: Option<String>,\n    pub merge_user_username: Option<String>,\n    pub created_at: i64,\n    pub updated_at: i64,\n    pub merged_at: Option<i64>,\n    pub closed_at: Option<i64>,\n    pub last_seen_at: i64,\n    pub web_url: String,\n}\n\n#[derive(Debug, Clone)]\npub struct MergeRequestWithMetadata {\n    pub merge_request: NormalizedMergeRequest,\n    pub label_names: Vec<String>,\n    pub assignee_usernames: Vec<String>,\n    pub reviewer_usernames: Vec<String>,\n}\n```\n\n## Function Signature\n```rust\npub fn transform_merge_request(\n    gitlab_mr: &GitLabMergeRequest,\n    local_project_id: i64,\n) -> Result<MergeRequestWithMetadata, String>\n```\n\n## Key Logic\n```rust\n// Draft: prefer draft, fallback to work_in_progress\nlet is_draft = gitlab_mr.draft || gitlab_mr.work_in_progress;\n\n// Merge status: prefer detailed_merge_status\nlet detailed_merge_status = gitlab_mr.detailed_merge_status\n    .clone()\n    .or_else(|| gitlab_mr.merge_status_legacy.clone());\n\n// Merge user: prefer merge_user\nlet merge_user_username = gitlab_mr.merge_user\n    .as_ref()\n    .map(|u| u.username.clone())\n    .or_else(|| gitlab_mr.merged_by.as_ref().map(|u| u.username.clone()));\n\n// References extraction\nlet (references_short, references_full) = gitlab_mr.references\n    .as_ref()\n    .map(|r| (Some(r.short.clone()), Some(r.full.clone())))\n    .unwrap_or((None, None));\n\n// Head SHA\nlet head_sha = gitlab_mr.sha.clone();\n```\n\n## Edge Cases\n- Invalid timestamps should return `Err`, not zero values\n- Empty labels/assignees/reviewers should return empty Vecs, not None\n- `state` must pass through as-is (including \"locked\")","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:40.849049Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:11:48.501301Z","closed_at":"2026-01-27T00:11:48.501241Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-34o","depends_on_id":"bd-3ir","type":"blocks","created_at":"2026-01-26T22:08:54.023616Z","created_by":"tayloreernisse"},{"issue_id":"bd-34o","depends_on_id":"bd-5ta","type":"blocks","created_at":"2026-01-26T22:08:54.059646Z","created_by":"tayloreernisse"}]}
 {"id":"bd-35o","title":"Create golden query test suite","description":"## Background\nGolden query tests verify end-to-end search quality with known-good expected results. They use a seeded SQLite DB with deterministic fixture data and fixed embedding vectors (no Ollama dependency). Each test query must return at least one expected URL in the top 10 results. These tests catch search regressions (ranking changes, filter bugs, missing results).\n\n## Approach\nCreate test infrastructure:\n\n**1. tests/fixtures/golden_queries.json:**\n```json\n[\n  {\n    \"query\": \"authentication login\",\n    \"mode\": \"lexical\",\n    \"filters\": {},\n    \"expected_urls\": [\"https://gitlab.example.com/group/project/-/issues/234\"],\n    \"min_results\": 1,\n    \"max_rank\": 10\n  },\n  {\n    \"query\": \"jwt token refresh\",\n    \"mode\": \"hybrid\",\n    \"filters\": {\"type\": \"merge_request\"},\n    \"expected_urls\": [\"https://gitlab.example.com/group/project/-/merge_requests/456\"],\n    \"min_results\": 1,\n    \"max_rank\": 10\n  }\n]\n```\n\n**2. Test harness (tests/golden_query_tests.rs):**\n- Load golden_queries.json\n- Create in-memory DB, apply all migrations\n- Seed with deterministic fixture documents (issues, MRs, discussions)\n- For hybrid/semantic queries: seed with fixed embedding vectors (768-dim, manually constructed for known similarity)\n- For each query: run search, verify expected URL in top N results\n\n**Fixture data design:**\n- 10-20 documents covering different source types\n- Known content that matches expected queries\n- Fixed embeddings: construct vectors where similar documents have small cosine distance\n- No randomness — fully deterministic\n\n## Acceptance Criteria\n- [ ] Golden queries file exists with at least 5 test queries\n- [ ] Test harness loads queries and validates each\n- [ ] All golden queries pass: expected URL in top 10\n- [ ] No external dependencies (no Ollama, no GitLab)\n- [ ] Deterministic fixture data (fixed embeddings, fixed content)\n- [ ] `cargo test --test golden_query_tests` passes in CI\n\n## Files\n- `tests/fixtures/golden_queries.json` — new file\n- `tests/golden_query_tests.rs` — new file (or tests/golden_queries.rs)\n\n## TDD Loop\nRED: Create golden_queries.json with expected results, harness fails (no fixture data)\nGREEN: Seed fixture data that satisfies expected results\nVERIFY: `cargo test --test golden_query_tests`\n\n## Edge Cases\n- Query matches multiple expected URLs: all must be present\n- Lexical queries: FTS ranking determines position, not vector\n- Hybrid queries: RRF combines both signals — fixed vectors must be designed to produce expected ranking\n- Empty result for a golden query: test failure with clear message showing actual results","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:27:21.788493Z","created_by":"tayloreernisse","updated_at":"2026-01-30T18:12:47.085563Z","closed_at":"2026-01-30T18:12:47.085363Z","close_reason":"Golden query test suite: 7 golden queries in fixture, 8 seeded documents, 2 test functions (all_pass + fixture_valid), deterministic in-memory DB, no external deps. 312 total tests pass.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-35o","depends_on_id":"bd-2no","type":"blocks","created_at":"2026-01-30T15:29:35.641568Z","created_by":"tayloreernisse"}]}
@@ -70,10 +92,12 @@
 {"id":"bd-38q","title":"Implement dirty source tracking module","description":"## Background\nDirty source tracking drives incremental document regeneration. When entities are upserted during ingestion, they're marked dirty. The regenerator drains this queue. The key constraint: mark_dirty_tx() takes &Transaction to enforce atomic marking inside the entity upsert transaction. Uses ON CONFLICT DO UPDATE (not INSERT OR IGNORE) to reset backoff on re-queue.\n\n## Approach\nCreate \\`src/ingestion/dirty_tracker.rs\\` per PRD Section 6.1.\n\n```rust\nconst DIRTY_SOURCES_BATCH_SIZE: usize = 500;\n\n/// Mark dirty INSIDE existing transaction. Takes &Transaction, NOT &Connection.\n/// ON CONFLICT resets ALL backoff/error state (not INSERT OR IGNORE).\n/// This ensures fresh updates are immediately eligible, not stuck behind stale backoff.\npub fn mark_dirty_tx(\n    tx: &rusqlite::Transaction<'_>,\n    source_type: SourceType,\n    source_id: i64,\n) -> Result<()>;\n\n/// Convenience wrapper for non-transactional contexts.\npub fn mark_dirty(conn: &Connection, source_type: SourceType, source_id: i64) -> Result<()>;\n\n/// Get dirty sources ready for processing.\n/// WHERE next_attempt_at IS NULL OR next_attempt_at <= now\n/// ORDER BY attempt_count ASC, queued_at ASC  (failed items deprioritized)\n/// LIMIT 500\npub fn get_dirty_sources(conn: &Connection) -> Result<Vec<(SourceType, i64)>>;\n\n/// Clear dirty entry after successful processing.\npub fn clear_dirty(conn: &Connection, source_type: SourceType, source_id: i64) -> Result<()>;\n```\n\n**PRD-specific details:**\n- get_dirty_sources ORDER BY: \\`attempt_count ASC, queued_at ASC\\` (failed items processed AFTER fresh items)\n- mark_dirty_tx ON CONFLICT resets: queued_at, attempt_count=0, last_attempt_at=NULL, last_error=NULL, next_attempt_at=NULL\n- SourceType parsed from string in query results via match on \\\"issue\\\"/\\\"merge_request\\\"/\\\"discussion\\\"\n- Invalid source_type in DB -> rusqlite::Error::FromSqlConversionFailure\n\n**Error recording is in regenerator.rs (bd-1u1)**, not dirty_tracker. The dirty_tracker only marks, gets, and clears.\n\n## Acceptance Criteria\n- [ ] mark_dirty_tx takes &Transaction<'_>, NOT &Connection\n- [ ] ON CONFLICT DO UPDATE resets: attempt_count=0, next_attempt_at=NULL, last_error=NULL, last_attempt_at=NULL\n- [ ] Uses ON CONFLICT DO UPDATE, NOT INSERT OR IGNORE (PRD explains why)\n- [ ] get_dirty_sources WHERE next_attempt_at IS NULL OR <= now\n- [ ] get_dirty_sources ORDER BY attempt_count ASC, queued_at ASC\n- [ ] get_dirty_sources LIMIT 500\n- [ ] get_dirty_sources returns Vec<(SourceType, i64)>\n- [ ] clear_dirty DELETEs entry\n- [ ] Queue drains completely when called in loop\n- [ ] \\`cargo test dirty_tracker\\` passes\n\n## Files\n- \\`src/ingestion/dirty_tracker.rs\\` — new file\n- \\`src/ingestion/mod.rs\\` — add \\`pub mod dirty_tracker;\\`\n\n## TDD Loop\nRED: Tests:\n- \\`test_mark_dirty_tx_inserts\\` — entry appears in dirty_sources\n- \\`test_requeue_resets_backoff\\` — mark, simulate error state, re-mark -> attempt_count=0, next_attempt_at=NULL\n- \\`test_get_respects_backoff\\` — entry with future next_attempt_at not returned\n- \\`test_get_orders_by_attempt_count\\` — fresh items before failed items\n- \\`test_batch_size_500\\` — insert 600, get returns 500\n- \\`test_clear_removes\\` — entry gone after clear\n- \\`test_drain_loop\\` — insert 1200, loop 3 times = empty\nGREEN: Implement all functions\nVERIFY: \\`cargo test dirty_tracker\\`\n\n## Edge Cases\n- Empty queue: get returns empty Vec\n- Invalid source_type string in DB: FromSqlConversionFailure error\n- Concurrent mark + get: ON CONFLICT handles race condition","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:27:09.434845Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:31:35.455315Z","closed_at":"2026-01-30T17:31:35.455127Z","close_reason":"Implemented dirty_tracker with mark_dirty_tx, get_dirty_sources, clear_dirty, record_dirty_error + 8 tests","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-38q","depends_on_id":"bd-36p","type":"blocks","created_at":"2026-01-30T15:29:34.914038Z","created_by":"tayloreernisse"},{"issue_id":"bd-38q","depends_on_id":"bd-hrs","type":"blocks","created_at":"2026-01-30T15:29:34.961390Z","created_by":"tayloreernisse"},{"issue_id":"bd-38q","depends_on_id":"bd-mem","type":"blocks","created_at":"2026-01-30T15:29:34.995197Z","created_by":"tayloreernisse"}]}
 {"id":"bd-39w","title":"[CP1] Test fixtures for mocked GitLab responses","description":"## Background\n\nTest fixtures provide mocked GitLab API responses for unit and integration tests. They enable testing without a live GitLab instance and ensure consistent test data across runs.\n\n## Approach\n\n### Fixture Files\n\nCreate JSON fixtures that match GitLab API response shapes:\n\n```\ntests/fixtures/\n├── gitlab_issue.json           # Single issue\n├── gitlab_issues_page.json     # Array of issues (pagination test)\n├── gitlab_discussion.json      # Single discussion with notes\n└── gitlab_discussions_page.json # Array of discussions\n```\n\n### gitlab_issue.json\n\n```json\n{\n  \"id\": 12345,\n  \"iid\": 42,\n  \"project_id\": 100,\n  \"title\": \"Test issue title\",\n  \"description\": \"Test issue description\",\n  \"state\": \"opened\",\n  \"created_at\": \"2024-01-15T10:00:00.000Z\",\n  \"updated_at\": \"2024-01-20T15:30:00.000Z\",\n  \"closed_at\": null,\n  \"author\": {\n    \"id\": 1,\n    \"username\": \"testuser\",\n    \"name\": \"Test User\"\n  },\n  \"labels\": [\"bug\", \"priority::high\"],\n  \"web_url\": \"https://gitlab.example.com/group/project/-/issues/42\"\n}\n```\n\n### gitlab_discussion.json\n\n```json\n{\n  \"id\": \"6a9c1750b37d513a43987b574953fceb50b03ce7\",\n  \"individual_note\": false,\n  \"notes\": [\n    {\n      \"id\": 1001,\n      \"type\": \"DiscussionNote\",\n      \"body\": \"First comment in thread\",\n      \"author\": { \"id\": 1, \"username\": \"testuser\", \"name\": \"Test User\" },\n      \"created_at\": \"2024-01-16T09:00:00.000Z\",\n      \"updated_at\": \"2024-01-16T09:00:00.000Z\",\n      \"system\": false,\n      \"resolvable\": true,\n      \"resolved\": false,\n      \"resolved_by\": null,\n      \"resolved_at\": null,\n      \"position\": null\n    },\n    {\n      \"id\": 1002,\n      \"type\": \"DiscussionNote\",\n      \"body\": \"Reply to first comment\",\n      \"author\": { \"id\": 2, \"username\": \"reviewer\", \"name\": \"Reviewer\" },\n      \"created_at\": \"2024-01-16T10:00:00.000Z\",\n      \"updated_at\": \"2024-01-16T10:00:00.000Z\",\n      \"system\": false,\n      \"resolvable\": true,\n      \"resolved\": false,\n      \"resolved_by\": null,\n      \"resolved_at\": null,\n      \"position\": null\n    }\n  ]\n}\n```\n\n### Helper Module\n\n```rust\n// tests/fixtures/mod.rs\n\npub fn load_fixture<T: DeserializeOwned>(name: &str) -> T {\n    let path = PathBuf::from(env!(\"CARGO_MANIFEST_DIR\"))\n        .join(\"tests/fixtures\")\n        .join(name);\n    let content = std::fs::read_to_string(&path)\n        .expect(&format!(\"Failed to read fixture: {}\", name));\n    serde_json::from_str(&content)\n        .expect(&format!(\"Failed to parse fixture: {}\", name))\n}\n\npub fn gitlab_issue() -> GitLabIssue {\n    load_fixture(\"gitlab_issue.json\")\n}\n\npub fn gitlab_issues_page() -> Vec<GitLabIssue> {\n    load_fixture(\"gitlab_issues_page.json\")\n}\n\npub fn gitlab_discussion() -> GitLabDiscussion {\n    load_fixture(\"gitlab_discussion.json\")\n}\n```\n\n## Acceptance Criteria\n\n- [ ] gitlab_issue.json deserializes to GitLabIssue correctly\n- [ ] gitlab_issues_page.json contains 3+ issues for pagination tests\n- [ ] gitlab_discussion.json contains multi-note thread\n- [ ] gitlab_discussions_page.json contains mix of individual_note true/false\n- [ ] At least one fixture includes system: true note\n- [ ] Helper functions load fixtures without panic\n\n## Files\n\n- tests/fixtures/gitlab_issue.json (create)\n- tests/fixtures/gitlab_issues_page.json (create)\n- tests/fixtures/gitlab_discussion.json (create)\n- tests/fixtures/gitlab_discussions_page.json (create)\n- tests/fixtures/mod.rs (create)\n\n## TDD Loop\n\nRED:\n```rust\n#[test] fn fixture_gitlab_issue_deserializes()\n#[test] fn fixture_gitlab_discussion_deserializes()\n#[test] fn fixture_has_system_note()\n```\n\nGREEN: Create JSON fixtures and helper module\n\nVERIFY: `cargo test fixture`\n\n## Edge Cases\n\n- Include issue with empty labels array\n- Include issue with null description\n- Include system note (system: true)\n- Include individual_note: true discussion (standalone comment)\n- Timestamps must be valid ISO 8601","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-25T17:02:38.433752Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:48:08.415195Z","closed_at":"2026-01-25T22:48:08.415132Z","close_reason":"Created 4 JSON fixture files (issue, issues_page, discussion, discussions_page) with helper tests - 6 tests passing","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-39w","depends_on_id":"bd-1np","type":"blocks","created_at":"2026-01-25T17:04:05.770848Z","created_by":"tayloreernisse"}]}
 {"id":"bd-3ae","title":"Epic: CP2 Gate A - MRs Only","description":"## Background\nGate A validates core MR ingestion works before adding complexity. Proves the cursor-based sync, pagination, and basic CLI work. This is the foundation - if Gate A fails, nothing else matters.\n\n## Acceptance Criteria (Pass/Fail)\n- [ ] `gi ingest --type=merge_requests` completes without error\n- [ ] `SELECT COUNT(*) FROM merge_requests` > 0\n- [ ] `gi list mrs --limit=5` shows 5 MRs with iid, title, state, author\n- [ ] `gi count mrs` shows total count matching DB query\n- [ ] MR with `state=locked` can be stored (if exists in test data)\n- [ ] Draft MR shows `draft=1` in DB and `[DRAFT]` in list output\n- [ ] `work_in_progress=true` MR shows `draft=1` (fallback works)\n- [ ] `head_sha` populated for MRs with commits\n- [ ] `references_short` and `references_full` populated\n- [ ] Re-run ingest shows \"0 new MRs\" or minimal refetch (cursor working)\n- [ ] Cursor saved at page boundary, not item boundary\n\n## Validation Script\n```bash\n#!/bin/bash\nset -e\n\nDB_PATH=\"${XDG_DATA_HOME:-$HOME/.local/share}/gitlab-inbox/db.sqlite3\"\n\necho \"=== Gate A: MRs Only ===\"\n\n# 1. Clear any existing MR data for clean test\necho \"Step 1: Reset MR cursor for clean test...\"\nsqlite3 \"$DB_PATH\" \"DELETE FROM sync_cursors WHERE resource_type = 'merge_requests';\"\n\n# 2. Run MR ingestion\necho \"Step 2: Ingest MRs...\"\ngi ingest --type=merge_requests\n\n# 3. Verify MRs exist\necho \"Step 3: Verify MR count...\"\nMR_COUNT=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM merge_requests;\")\necho \"  MR count: $MR_COUNT\"\n[ \"$MR_COUNT\" -gt 0 ] || { echo \"FAIL: No MRs ingested\"; exit 1; }\n\n# 4. Verify list command\necho \"Step 4: Test list command...\"\ngi list mrs --limit=5\n\n# 5. Verify count command\necho \"Step 5: Test count command...\"\ngi count mrs\n\n# 6. Verify draft handling\necho \"Step 6: Check draft MRs...\"\nDRAFT_COUNT=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM merge_requests WHERE draft = 1;\")\necho \"  Draft MR count: $DRAFT_COUNT\"\n\n# 7. Verify head_sha population\necho \"Step 7: Check head_sha...\"\nSHA_COUNT=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM merge_requests WHERE head_sha IS NOT NULL;\")\necho \"  MRs with head_sha: $SHA_COUNT\"\n\n# 8. Verify references\necho \"Step 8: Check references...\"\nREF_COUNT=$(sqlite3 \"$DB_PATH\" \"SELECT COUNT(*) FROM merge_requests WHERE references_short IS NOT NULL;\")\necho \"  MRs with references: $REF_COUNT\"\n\n# 9. Verify cursor saved\necho \"Step 9: Check cursor...\"\nCURSOR=$(sqlite3 \"$DB_PATH\" \"SELECT updated_at, gitlab_id FROM sync_cursors WHERE resource_type = 'merge_requests';\")\necho \"  Cursor: $CURSOR\"\n[ -n \"$CURSOR\" ] || { echo \"FAIL: Cursor not saved\"; exit 1; }\n\n# 10. Re-run and verify minimal refetch\necho \"Step 10: Re-run ingest (should be minimal)...\"\ngi ingest --type=merge_requests\n# Output should show minimal or zero new MRs\n\necho \"\"\necho \"=== Gate A: PASSED ===\"\n```\n\n## Test Commands (Quick Verification)\n```bash\n# Run these in order:\ngi ingest --type=merge_requests\ngi list mrs --limit=10\ngi count mrs\n\n# Verify in DB:\nsqlite3 ~/.local/share/gitlab-inbox/db.sqlite3 \"\n  SELECT \n    COUNT(*) as total,\n    SUM(CASE WHEN draft = 1 THEN 1 ELSE 0 END) as drafts,\n    SUM(CASE WHEN head_sha IS NOT NULL THEN 1 ELSE 0 END) as with_sha,\n    SUM(CASE WHEN references_short IS NOT NULL THEN 1 ELSE 0 END) as with_refs\n  FROM merge_requests;\n\"\n\n# Re-run (should be no-op):\ngi ingest --type=merge_requests\n```\n\n## Dependencies\nThis gate requires these beads to be complete:\n- bd-3ir (Database migration)\n- bd-5ta (GitLab MR types)\n- bd-34o (MR transformer)\n- bd-iba (GitLab client pagination)\n- bd-ser (MR ingestion module)\n\n## Edge Cases\n- `locked` state is transitional (merge in progress); may not exist in test data\n- Some older GitLab instances may not return `head_sha` for all MRs\n- `work_in_progress` is deprecated but should still work as fallback\n- Very large projects (10k+ MRs) may take significant time on first sync","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-26T22:06:00.966522Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:48:21.057298Z","closed_at":"2026-01-27T00:48:21.057225Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-3ae","depends_on_id":"bd-iba","type":"blocks","created_at":"2026-01-26T22:08:55.576626Z","created_by":"tayloreernisse"},{"issue_id":"bd-3ae","depends_on_id":"bd-ser","type":"blocks","created_at":"2026-01-26T22:08:55.446814Z","created_by":"tayloreernisse"}]}
+{"id":"bd-3as","title":"Implement timeline event collection and chronological interleaving","description":"## Background\nSteps 4-5 of timeline query: collect events from all sources for seed + expanded entities and sort chronologically. This produces the final event stream for rendering.\n\n## Approach\nAdd to src/core/timeline.rs:\n\n```rust\n/// Collect all events for the given entities and sort chronologically.\npub fn collect_and_interleave_events(\n    conn: &Connection,\n    seeds: &[SeedEntity],\n    expanded: &[ExpandedEntity],\n    evidence_notes: &[EvidenceNote],\n    limit: Option<usize>,          // -n flag\n) -> Result<Vec<TimelineEvent>>\n```\n\nFor each entity (seed + expanded), collect:\n\n1. **Entity creation** — from issues/merge_requests created_at:\n```sql\nSELECT created_at, title, author_username, web_url FROM issues WHERE id = ?;\n```\n→ TimelineEvent { event_type: Created, ... }\n\n2. **State changes** — from resource_state_events:\n```sql\nSELECT state, actor_username, created_at FROM resource_state_events\nWHERE (issue_id = ? OR merge_request_id = ?) ORDER BY created_at;\n```\n→ TimelineEvent { event_type: StateChanged { state }, ... }\n\n3. **Label changes** — from resource_label_events:\n```sql\nSELECT label_name, action, actor_username, created_at FROM resource_label_events\nWHERE (issue_id = ? OR merge_request_id = ?) ORDER BY created_at;\n```\n→ TimelineEvent { event_type: LabelAdded/LabelRemoved { label }, ... }\n\n4. **Milestone changes** — from resource_milestone_events:\n```sql\nSELECT milestone_title, action, actor_username, created_at FROM resource_milestone_events\nWHERE (issue_id = ? OR merge_request_id = ?) ORDER BY created_at;\n```\n→ TimelineEvent { event_type: MilestoneSet/MilestoneRemoved { milestone }, ... }\n\n5. **Merge events** — from merge_requests.merged_at (if not NULL):\n```sql\nSELECT merged_at, author_username, web_url FROM merge_requests WHERE id = ? AND merged_at IS NOT NULL;\n```\n→ TimelineEvent { event_type: Merged, ... }\n\n6. **Evidence notes** — from the seed phase's evidence_notes list:\n→ TimelineEvent { event_type: NoteEvidence { note_id, snippet, discussion_id }, ... }\n\nMark each event with is_seed based on whether the entity is in seeds vs expanded.\n\nSort all events using TimelineEvent's Ord implementation (chronological + stable tiebreak).\n\nApply -n limit after sorting.\n\n## Acceptance Criteria\n- [ ] All 6 event types collected correctly\n- [ ] Entity creation events have correct timestamps and metadata\n- [ ] State/label/milestone events from resource event tables\n- [ ] Merge events only for merged MRs\n- [ ] Evidence notes converted to NoteEvidence events with truncated snippets (~200 chars)\n- [ ] is_seed flag correct (seed entities = true, expanded = false)\n- [ ] Chronological ordering with stable tiebreak\n- [ ] -n limit applied after sorting\n\n## Files\n- src/core/timeline.rs (add collect_and_interleave_events)\n\n## TDD Loop\nRED: tests/timeline_collect_tests.rs:\n- `test_collect_entity_creation` - verify Created event for an issue\n- `test_collect_state_changes` - seed state events, verify StateChanged events\n- `test_collect_label_changes` - seed label events, verify LabelAdded/Removed\n- `test_collect_merge_events` - merged MR produces Merged event\n- `test_collect_evidence_notes` - evidence notes become NoteEvidence events\n- `test_collect_chronological_order` - events sorted by timestamp\n- `test_collect_is_seed_flag` - seed entity events = true, expanded = false\n- `test_collect_limit` - 20 events, limit 5, only 5 returned\n\nSetup: create_test_db with full migrations, seed entities + events.\n\nGREEN: Implement collection queries and sorting\n\nVERIFY: `cargo test timeline_collect -- --nocapture`\n\n## Edge Cases\n- Entity with no events (newly created, no state changes yet) — only Created event produced\n- Evidence note snippet truncation: handle UTF-8 boundary correctly (use char_indices)\n- Merge event may have NULL actor (auto-merge) — actor is Optional\n- resource event tables may be empty (user hasn't synced events) — graceful, just no event entries\n- Very large entity sets (100+ entities × 50+ events each) — ensure efficient batching of queries","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:08.703942Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:46:32.854764Z","compaction_level":0,"original_size":0,"labels":["gate-3","phase-b","query"],"dependencies":[{"issue_id":"bd-3as","depends_on_id":"bd-1ep","type":"blocks","created_at":"2026-02-02T21:33:37.618171Z","created_by":"tayloreernisse"},{"issue_id":"bd-3as","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:08.705605Z","created_by":"tayloreernisse"},{"issue_id":"bd-3as","depends_on_id":"bd-ypa","type":"blocks","created_at":"2026-02-02T21:33:37.575585Z","created_by":"tayloreernisse"}]}
 {"id":"bd-3bo","title":"[CP1] gi count issues/discussions/notes commands","description":"Count entities in the database.\n\nCommands:\n- gi count issues → 'Issues: N'\n- gi count discussions --type=issue → 'Issue Discussions: N'\n- gi count notes --type=issue → 'Issue Notes: N (excluding M system)'\n\nFiles: src/cli/commands/count.ts\nDone when: Counts match expected values from GitLab","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T15:20:16.190875Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.156293Z","deleted_at":"2026-01-25T15:21:35.156290Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-3eu","title":"Implement hybrid search with adaptive recall","description":"## Background\nHybrid search is the top-level search orchestrator that combines FTS5 lexical results with sqlite-vec semantic results via RRF ranking. It supports three modes (Lexical, Semantic, Hybrid) and implements adaptive recall (wider initial fetch when filters are applied) and graceful degradation (falls back to FTS when Ollama is unavailable). All modes use RRF for consistent --explain output.\n\n## Approach\nCreate `src/search/hybrid.rs` per PRD Section 5.3.\n\n**Key types:**\n```rust\n#[derive(Debug, Clone, Copy, PartialEq, Eq)]\npub enum SearchMode {\n    Hybrid,    // Vector + FTS with RRF\n    Lexical,   // FTS only\n    Semantic,  // Vector only\n}\n\nimpl SearchMode {\n    pub fn from_str(s: &str) -> Option<Self> {\n        match s.to_lowercase().as_str() {\n            \"hybrid\" => Some(Self::Hybrid),\n            \"lexical\" | \"fts\" => Some(Self::Lexical),\n            \"semantic\" | \"vector\" => Some(Self::Semantic),\n            _ => None,\n        }\n    }\n\n    pub fn as_str(&self) -> &'static str {\n        match self {\n            Self::Hybrid => \"hybrid\",\n            Self::Lexical => \"lexical\",\n            Self::Semantic => \"semantic\",\n        }\n    }\n}\n\npub struct HybridResult {\n    pub document_id: i64,\n    pub score: f64,          // Normalized RRF score (0-1)\n    pub vector_rank: Option<usize>,\n    pub fts_rank: Option<usize>,\n    pub rrf_score: f64,      // Raw RRF score\n}\n```\n\n**Core function (ASYNC, PRD-exact signature):**\n```rust\npub async fn search_hybrid(\n    conn: &Connection,\n    client: Option<&OllamaClient>,      // None if Ollama unavailable\n    ollama_base_url: Option<&str>,       // For actionable error messages\n    query: &str,\n    mode: SearchMode,\n    filters: &SearchFilters,\n    fts_mode: FtsQueryMode,\n) -> Result<(Vec<HybridResult>, Vec<String>)>\n```\n\n**IMPORTANT — client is `Option<&OllamaClient>`:** This enables graceful degradation. When Ollama is unavailable, the caller passes `None` and hybrid mode falls back to FTS-only with a warning. The `ollama_base_url` is separate so error messages can include it even when client is None.\n\n**Adaptive recall constants (PRD Section 5.3):**\n```rust\nconst BASE_RECALL_MIN: usize = 50;\nconst FILTERED_RECALL_MIN: usize = 200;\nconst RECALL_CAP: usize = 1500;\n```\n\n**Recall formula:**\n```rust\nlet requested = filters.clamp_limit();\nlet top_k = if filters.has_any_filter() {\n    (requested * 50).max(FILTERED_RECALL_MIN).min(RECALL_CAP)\n} else {\n    (requested * 10).max(BASE_RECALL_MIN).min(RECALL_CAP)\n};\n```\n\n**Mode behavior:**\n- **Lexical:** FTS only -> rank_rrf with empty vector list (single-list RRF)\n- **Semantic:** Vector only -> requires client (error if None) -> rank_rrf with empty FTS list\n- **Hybrid:** Both FTS + vector -> rank_rrf with both lists\n- **Hybrid with client=None:** Graceful degradation to Lexical with warning, NOT error\n\n**Graceful degradation logic:**\n```rust\nSearchMode::Hybrid => {\n    let fts_results = search_fts(conn, query, top_k, fts_mode)?;\n    let fts_tuples: Vec<_> = fts_results.iter().map(|r| (r.document_id, r.rank)).collect();\n\n    match client {\n        Some(client) => {\n            let query_embedding = client.embed_batch(vec\\![query.to_string()]).await?;\n            let embedding = query_embedding.into_iter().next().unwrap();\n            let vec_results = search_vector(conn, &embedding, top_k)?;\n            let vec_tuples: Vec<_> = vec_results.iter().map(|r| (r.document_id, r.distance)).collect();\n            let ranked = rank_rrf(&vec_tuples, &fts_tuples);\n            // ... map to HybridResult\n            Ok((results, warnings))\n        }\n        None => {\n            warnings.push(\"Ollama unavailable, falling back to lexical search\".into());\n            let ranked = rank_rrf(&[], &fts_tuples);\n            // ... map to HybridResult\n            Ok((results, warnings))\n        }\n    }\n}\n```\n\n## Acceptance Criteria\n- [ ] Function is `async` (per PRD — Ollama client methods are async)\n- [ ] Signature takes `client: Option<&OllamaClient>` (not required)\n- [ ] Signature takes `ollama_base_url: Option<&str>` for actionable error messages\n- [ ] Returns `(Vec<HybridResult>, Vec<String>)` — results + warnings\n- [ ] Lexical mode: FTS-only results ranked via RRF (single list)\n- [ ] Semantic mode: vector-only results ranked via RRF; error if client is None\n- [ ] Hybrid mode: both FTS + vector results merged via RRF\n- [ ] Graceful degradation: client=None in Hybrid falls back to FTS with warning (not error)\n- [ ] Adaptive recall: unfiltered max(50, limit*10), filtered max(200, limit*50), capped 1500\n- [ ] All modes produce consistent --explain output (vector_rank, fts_rank, rrf_score)\n- [ ] SearchMode::from_str accepts aliases: \"fts\" for Lexical, \"vector\" for Semantic\n- [ ] `cargo build` succeeds\n\n## Files\n- `src/search/hybrid.rs` — new file\n- `src/search/mod.rs` — add `pub use hybrid::{search_hybrid, HybridResult, SearchMode};`\n\n## TDD Loop\nRED: Tests (some integration, some unit):\n- `test_lexical_mode` — FTS results only\n- `test_semantic_mode` — vector results only\n- `test_hybrid_mode` — both lists merged\n- `test_graceful_degradation` — None client falls back to FTS with warning in warnings vec\n- `test_adaptive_recall_unfiltered` — recall = max(50, limit*10)\n- `test_adaptive_recall_filtered` — recall = max(200, limit*50)\n- `test_recall_cap` — never exceeds 1500\n- `test_search_mode_from_str` — \"hybrid\", \"lexical\", \"fts\", \"semantic\", \"vector\", invalid\nGREEN: Implement search_hybrid\nVERIFY: `cargo test hybrid`\n\n## Edge Cases\n- Both FTS and vector return zero results: empty output (not error)\n- FTS returns results but vector returns empty: RRF still works (single-list)\n- Very high limit (100) with filters: recall = min(5000, 1500) = 1500\n- Semantic mode with client=None: error (OllamaUnavailable), not degradation\n- Semantic mode with 0% coverage: return LoreError::EmbeddingsNotBuilt","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:50.343002Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:56:16.631748Z","closed_at":"2026-01-30T17:56:16.631682Z","close_reason":"Implemented hybrid search with 3 modes (lexical/semantic/hybrid), graceful degradation when Ollama unavailable, adaptive recall (50-1500), RRF fusion. 6 tests pass.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-3eu","depends_on_id":"bd-1k1","type":"blocks","created_at":"2026-01-30T15:29:24.913458Z","created_by":"tayloreernisse"},{"issue_id":"bd-3eu","depends_on_id":"bd-335","type":"blocks","created_at":"2026-01-30T15:29:25.025502Z","created_by":"tayloreernisse"},{"issue_id":"bd-3eu","depends_on_id":"bd-3ez","type":"blocks","created_at":"2026-01-30T15:29:24.987809Z","created_by":"tayloreernisse"},{"issue_id":"bd-3eu","depends_on_id":"bd-bjo","type":"blocks","created_at":"2026-01-30T15:29:24.950761Z","created_by":"tayloreernisse"}]}
 {"id":"bd-3ez","title":"Implement RRF ranking","description":"## Background\nReciprocal Rank Fusion (RRF) combines results from multiple retrieval systems (FTS5 lexical + sqlite-vec semantic) into a single ranked list without requiring score normalization. Documents appearing in both lists rank higher than single-list documents. This is the core ranking algorithm for hybrid search in Gate B.\n\n## Approach\nCreate \\`src/search/rrf.rs\\` per PRD Section 5.2.\n\n```rust\nuse std::collections::HashMap;\n\nconst RRF_K: f64 = 60.0;\n\npub struct RrfResult {\n    pub document_id: i64,\n    pub rrf_score: f64,           // Raw RRF score\n    pub normalized_score: f64,    // Normalized to 0-1 (rrf_score / max)\n    pub vector_rank: Option<usize>,  // 1-indexed rank in vector list\n    pub fts_rank: Option<usize>,     // 1-indexed rank in FTS list\n}\n\n/// Input: tuples of (document_id, score/distance) — already sorted by retriever.\n/// Ranks are 1-indexed (first result = rank 1).\n/// Score = sum of 1/(k + rank) for each list containing the document.\npub fn rank_rrf(\n    vector_results: &[(i64, f64)],  // (doc_id, distance)\n    fts_results: &[(i64, f64)],     // (doc_id, bm25_score)\n) -> Vec<RrfResult>\n```\n\n**Algorithm (per PRD):**\n1. Build HashMap<doc_id, (rrf_score, vector_rank, fts_rank)>\n2. For each vector result at position i: score += 1/(K + (i+1)), record vector_rank = i+1 (**1-indexed**)\n3. For each FTS result at position i: score += 1/(K + (i+1)), record fts_rank = i+1 (**1-indexed**)\n4. Sort descending by rrf_score\n5. Normalize: each result.normalized_score = result.rrf_score / max_score (best = 1.0)\n\n**Key PRD details:**\n- Ranks are **1-indexed** (rank 1 = best, not rank 0)\n- Input is \\`&[(i64, f64)]\\` tuples, NOT custom structs\n- Output has both \\`rrf_score\\` (raw) and \\`normalized_score\\` (0-1)\n\n## Acceptance Criteria\n- [ ] Documents in both lists score higher than single-list documents\n- [ ] Single-list documents are included (not dropped)\n- [ ] Ranks are 1-indexed (first element = rank 1)\n- [ ] Raw RRF score available in rrf_score field\n- [ ] Normalized score: best = 1.0, all in [0, 1]\n- [ ] Results sorted descending by rrf_score\n- [ ] vector_rank and fts_rank tracked per result for --explain\n- [ ] Empty input lists handled (return empty)\n- [ ] One empty list + one non-empty returns results from non-empty list\n\n## Files\n- \\`src/search/rrf.rs\\` — new file\n- \\`src/search/mod.rs\\` — add \\`mod rrf; pub use rrf::{rank_rrf, RrfResult};\\`\n\n## TDD Loop\nRED: Tests in \\`#[cfg(test)] mod tests\\`:\n- \\`test_dual_list_ranks_higher\\` — doc in both lists scores > doc in one list\n- \\`test_single_list_included\\` — FTS-only and vector-only docs appear\n- \\`test_normalization\\` — best score is 1.0, all in [0, 1]\n- \\`test_empty_inputs\\` — empty returns empty\n- \\`test_ranks_are_1_indexed\\` — verify vector_rank/fts_rank start at 1\n- \\`test_raw_and_normalized_scores\\` — both fields populated correctly\nGREEN: Implement rank_rrf()\nVERIFY: \\`cargo test rrf\\`\n\n## Edge Cases\n- Duplicate document_id within same list: shouldn't happen, use first occurrence\n- Single result in one list, zero in other: normalized_score = 1.0\n- Very large input lists: HashMap handles efficiently","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:50.309012Z","created_by":"tayloreernisse","updated_at":"2026-01-30T16:53:04.128560Z","closed_at":"2026-01-30T16:53:04.128498Z","close_reason":"Completed: RRF ranking with 1-indexed ranks, raw+normalized scores, vector_rank/fts_rank provenance, 7 tests pass","compaction_level":0,"original_size":0}
 {"id":"bd-3hy","title":"[CP1] Test fixtures for mocked GitLab responses","description":"Create mock response files for integration tests using wiremock.\n\n## Fixtures to Create\n\n### tests/fixtures/gitlab_issue.json\nSingle issue with labels:\n- id, iid, project_id, title, description, state\n- author object\n- labels array (string names)\n- timestamps\n- web_url\n\n### tests/fixtures/gitlab_issues_page.json\nArray of issues simulating paginated response:\n- 3-5 issues with varying states\n- Mix of labels\n\n### tests/fixtures/gitlab_discussion.json\nSingle discussion:\n- id (string)\n- individual_note: false\n- notes array with 2+ notes\n- Include one system note\n\n### tests/fixtures/gitlab_discussions_page.json\nArray of discussions:\n- Mix of individual_note true/false\n- Include resolvable/resolved examples\n\n## Edge Cases to Cover\n- Issue with no labels (empty array)\n- Issue with labels_details (ignored in CP1)\n- Discussion with individual_note=true (single note)\n- System notes with system=true\n- Resolvable notes\n\nFiles: tests/fixtures/gitlab_issue.json, gitlab_issues_page.json, gitlab_discussion.json, gitlab_discussions_page.json\nDone when: wiremock handlers can use fixtures for deterministic tests","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:59:01.206436Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.991367Z","deleted_at":"2026-01-25T17:02:01.991362Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-3ia","title":"Fetch closes_issues API and populate entity_references","description":"## Background\nGET /projects/:id/merge_requests/:iid/closes_issues returns issues that will close when MR merges. This is the most reliable source for MR→issue relationships. Uses the generic dependent fetch queue (job_type = 'mr_closes_issues').\n\n## Approach\n\n**1. Add API endpoint to GitLab client (src/gitlab/client.rs):**\n```rust\n/// Fetch issues that will be closed when this MR merges.\npub async fn fetch_mr_closes_issues(\n    &self, \n    project_id: i64, \n    iid: i64\n) -> Result<Vec<GitLabIssueRef>>\n```\n\nNew type in src/gitlab/types.rs:\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabIssueRef {\n    pub id: i64,\n    pub iid: i64,\n    pub project_id: i64,\n    pub title: String,\n    pub state: String,\n    pub web_url: String,\n}\n```\n\nURL: `GET /api/v4/projects/{project_id}/merge_requests/{iid}/closes_issues?per_page=100`\n\n**2. Enqueue jobs during MR ingestion:**\nIn orchestrator.rs, after MR upsert:\n```rust\nenqueue_job(conn, project_id, \"merge_request\", iid, local_id, \"mr_closes_issues\", None)?;\n```\n\nThis is always enqueued (not gated by a config flag) because cross-reference data is fundamental to all temporal queries.\n\n**3. Process jobs in drain step:**\nIn the drain dispatcher (from bd-1ep), handle \"mr_closes_issues\" job_type:\n```rust\nlet closes_issues = client.fetch_mr_closes_issues(gitlab_project_id, job.entity_iid).await?;\nfor issue_ref in &closes_issues {\n    let target_id = resolve_issue_local_id(conn, project_id, issue_ref.iid);\n    insert_entity_reference(conn, EntityReference {\n        source_entity_type: \"merge_request\",\n        source_entity_id: job.entity_local_id,\n        target_entity_type: \"issue\",\n        target_entity_id: target_id,            // Some(id) or None for cross-project\n        target_project_path: if target_id.is_none() { Some(resolve_project_path(issue_ref.project_id)) } else { None },\n        target_entity_iid: if target_id.is_none() { Some(issue_ref.iid) } else { None },\n        reference_type: \"closes\",\n        source_method: \"api_closes_issues\",\n        created_at: None,\n    })?;\n}\n```\n\n**4. Insert helper for entity_references:**\nAdd to src/core/references.rs:\n```rust\npub fn insert_entity_reference(conn: &Connection, ref_: &EntityReference) -> Result<bool>\n// INSERT OR IGNORE, returns true if inserted\n```\n\n## Acceptance Criteria\n- [ ] closes_issues API called for all MRs during sync\n- [ ] Entity references created with reference_type='closes', source_method='api_closes_issues'\n- [ ] Source = MR, target = issue (correct directionality)\n- [ ] Cross-project issues stored as unresolved (target_entity_id=NULL, target_project_path set)\n- [ ] Idempotent: re-sync doesn't create duplicate references\n- [ ] 404 on deleted MR handled gracefully (fail_job)\n\n## Files\n- src/gitlab/client.rs (add fetch_mr_closes_issues)\n- src/gitlab/types.rs (add GitLabIssueRef)\n- src/core/references.rs (add insert_entity_reference helper)\n- src/ingestion/orchestrator.rs (enqueue mr_closes_issues jobs)\n- src/core/drain.rs or sync.rs (handle mr_closes_issues in drain dispatcher)\n\n## TDD Loop\nRED: tests/references_tests.rs:\n- `test_closes_issues_creates_references` - mock closes_issues response, verify entity_references rows\n- `test_closes_issues_cross_project_unresolved` - issue from different project stored as unresolved\n- `test_closes_issues_idempotent` - process same job twice, verify no duplicates\n\ntests/gitlab_types_tests.rs:\n- `test_deserialize_issue_ref` - verify GitLabIssueRef deserialization\n\nGREEN: Implement API endpoint, enqueue hook, drain handler, insert helper\n\nVERIFY: `cargo test references -- --nocapture && cargo test gitlab_types -- --nocapture`\n\n## Edge Cases\n- closes_issues API returns issues from OTHER projects (cross-project closing) — must check if issue is in local DB\n- Empty response (MR doesn't close any issues) — no refs created, job still completed\n- MR may close the same issue via description (\"Closes #123\") and via commits — API deduplicates, but our INSERT OR IGNORE handles it too\n- The closes_issues API may return stale data for draft MRs (issues that *would* close but haven't yet)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:32:33.561956Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:41:50.613792Z","compaction_level":0,"original_size":0,"labels":["api","gate-2","phase-b"],"dependencies":[{"issue_id":"bd-3ia","depends_on_id":"bd-1se","type":"parent-child","created_at":"2026-02-02T21:32:33.563366Z","created_by":"tayloreernisse"},{"issue_id":"bd-3ia","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T22:41:50.613776Z","created_by":"tayloreernisse"},{"issue_id":"bd-3ia","depends_on_id":"bd-tir","type":"blocks","created_at":"2026-02-02T21:32:42.860463Z","created_by":"tayloreernisse"}]}
 {"id":"bd-3ir","title":"Add database migration 006_merge_requests.sql","description":"## Background\nFoundation for all CP2 MR features. This migration defines the schema that all other MR components depend on. Must complete BEFORE any other CP2 work can proceed.\n\n## Approach\nCreate migration file that adds:\n1. `merge_requests` table with all CP2 fields\n2. `mr_labels`, `mr_assignees`, `mr_reviewers` junction tables\n3. Indexes on discussions for MR queries\n4. DiffNote position columns on notes table\n\n## Files\n- `migrations/006_merge_requests.sql` - New migration file\n- `src/core/db.rs` - Update MIGRATIONS const to include version 6\n\n## Acceptance Criteria\n- [ ] Migration file exists at `migrations/006_merge_requests.sql`\n- [ ] `merge_requests` table has columns: id, gitlab_id, project_id, iid, title, description, state, draft, author_username, source_branch, target_branch, head_sha, references_short, references_full, detailed_merge_status, merge_user_username, created_at, updated_at, merged_at, closed_at, last_seen_at, discussions_synced_for_updated_at, discussions_sync_last_attempt_at, discussions_sync_attempts, discussions_sync_last_error, web_url, raw_payload_id\n- [ ] `mr_labels` junction table exists with (merge_request_id, label_id) PK\n- [ ] `mr_assignees` junction table exists with (merge_request_id, username) PK\n- [ ] `mr_reviewers` junction table exists with (merge_request_id, username) PK\n- [ ] `idx_discussions_mr_id` and `idx_discussions_mr_resolved` indexes exist\n- [ ] `notes` table has new columns: position_type, position_line_range_start, position_line_range_end, position_base_sha, position_start_sha, position_head_sha\n- [ ] `gi doctor` runs without migration errors\n- [ ] `cargo test` passes\n\n## TDD Loop\nRED: Cannot open DB with version 6 schema\nGREEN: Add migration file with full SQL\nVERIFY: `cargo run -- doctor` shows healthy DB\n\n## SQL Reference (from PRD)\n```sql\n-- Merge requests table\nCREATE TABLE merge_requests (\n  id INTEGER PRIMARY KEY,\n  gitlab_id INTEGER UNIQUE NOT NULL,\n  project_id INTEGER NOT NULL REFERENCES projects(id),\n  iid INTEGER NOT NULL,\n  title TEXT,\n  description TEXT,\n  state TEXT,                         -- opened | merged | closed | locked\n  draft INTEGER NOT NULL DEFAULT 0,   -- SQLite boolean\n  author_username TEXT,\n  source_branch TEXT,\n  target_branch TEXT,\n  head_sha TEXT,\n  references_short TEXT,\n  references_full TEXT,\n  detailed_merge_status TEXT,\n  merge_user_username TEXT,\n  created_at INTEGER,                 -- ms epoch UTC\n  updated_at INTEGER,\n  merged_at INTEGER,\n  closed_at INTEGER,\n  last_seen_at INTEGER NOT NULL,\n  discussions_synced_for_updated_at INTEGER,\n  discussions_sync_last_attempt_at INTEGER,\n  discussions_sync_attempts INTEGER DEFAULT 0,\n  discussions_sync_last_error TEXT,\n  web_url TEXT,\n  raw_payload_id INTEGER REFERENCES raw_payloads(id)\n);\nCREATE INDEX idx_mrs_project_updated ON merge_requests(project_id, updated_at);\nCREATE UNIQUE INDEX uq_mrs_project_iid ON merge_requests(project_id, iid);\n-- ... (see PRD for full index list)\n\n-- Junction tables\nCREATE TABLE mr_labels (\n  merge_request_id INTEGER REFERENCES merge_requests(id) ON DELETE CASCADE,\n  label_id INTEGER REFERENCES labels(id) ON DELETE CASCADE,\n  PRIMARY KEY(merge_request_id, label_id)\n);\n\nCREATE TABLE mr_assignees (\n  merge_request_id INTEGER REFERENCES merge_requests(id) ON DELETE CASCADE,\n  username TEXT NOT NULL,\n  PRIMARY KEY(merge_request_id, username)\n);\n\nCREATE TABLE mr_reviewers (\n  merge_request_id INTEGER REFERENCES merge_requests(id) ON DELETE CASCADE,\n  username TEXT NOT NULL,\n  PRIMARY KEY(merge_request_id, username)\n);\n\n-- DiffNote position columns (ALTER TABLE)\nALTER TABLE notes ADD COLUMN position_type TEXT;\nALTER TABLE notes ADD COLUMN position_line_range_start INTEGER;\nALTER TABLE notes ADD COLUMN position_line_range_end INTEGER;\nALTER TABLE notes ADD COLUMN position_base_sha TEXT;\nALTER TABLE notes ADD COLUMN position_start_sha TEXT;\nALTER TABLE notes ADD COLUMN position_head_sha TEXT;\n\nINSERT INTO schema_version (version, applied_at, description)\nVALUES (6, strftime('%s', 'now') * 1000, 'Merge requests, MR labels, assignees, reviewers');\n```\n\n## Edge Cases\n- SQLite does not support ADD CONSTRAINT - FK defined as nullable in CP1\n- `locked` state is transitional (merge-in-progress) - store as first-class\n- discussions_synced_for_updated_at prevents redundant discussion refetch","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:40.101470Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:06:43.899079Z","closed_at":"2026-01-27T00:06:43.898875Z","close_reason":"Migration 006_merge_requests.sql created and verified. Schema v6 applied successfully with all tables, indexes, and position columns.","compaction_level":0,"original_size":0}
 {"id":"bd-3j6","title":"Add transform_mr_discussion and transform_notes_with_diff_position","description":"## Background\nExtends discussion transformer for MR context. MR discussions can contain DiffNotes with file position metadata. This is critical for code review context in CP3 document generation.\n\n## Approach\nAdd two new functions to existing `src/gitlab/transformers/discussion.rs`:\n1. `transform_mr_discussion()` - Transform discussion with MR reference\n2. `transform_notes_with_diff_position()` - Extract DiffNote position metadata\n\nCP1 already has the polymorphic `NormalizedDiscussion` with `NoteableRef` enum - reuse that pattern.\n\n## Files\n- `src/gitlab/transformers/discussion.rs` - Add new functions\n- `tests/diffnote_tests.rs` - DiffNote position extraction tests\n- `tests/mr_discussion_tests.rs` - MR discussion transform tests\n\n## Acceptance Criteria\n- [ ] `transform_mr_discussion()` returns `NormalizedDiscussion` with `merge_request_id: Some(local_mr_id)`\n- [ ] `transform_notes_with_diff_position()` returns `Result<Vec<NormalizedNote>, String>`\n- [ ] DiffNote position fields extracted: `position_old_path`, `position_new_path`, `position_old_line`, `position_new_line`\n- [ ] Extended position fields extracted: `position_type`, `position_line_range_start`, `position_line_range_end`\n- [ ] SHA triplet extracted: `position_base_sha`, `position_start_sha`, `position_head_sha`\n- [ ] Strict timestamp parsing - returns `Err` on invalid timestamps (no `unwrap_or(0)`)\n- [ ] `cargo test diffnote` passes\n- [ ] `cargo test mr_discussion` passes\n\n## TDD Loop\nRED: `cargo test diffnote_position` -> test fails\nGREEN: Add position extraction logic\nVERIFY: `cargo test diffnote`\n\n## Function Signatures\n```rust\n/// Transform GitLab discussion for MR context.\n/// Reuses existing transform_discussion logic, just with MR reference.\npub fn transform_mr_discussion(\n    gitlab_discussion: &GitLabDiscussion,\n    local_project_id: i64,\n    local_mr_id: i64,\n) -> NormalizedDiscussion {\n    // Use existing transform_discussion with NoteableRef::MergeRequest(local_mr_id)\n    transform_discussion(\n        gitlab_discussion,\n        local_project_id,\n        NoteableRef::MergeRequest(local_mr_id),\n    )\n}\n\n/// Transform notes with DiffNote position extraction.\n/// Returns Result to enforce strict timestamp parsing.\npub fn transform_notes_with_diff_position(\n    gitlab_discussion: &GitLabDiscussion,\n    local_project_id: i64,\n) -> Result<Vec<NormalizedNote>, String>\n```\n\n## DiffNote Position Extraction\n```rust\n// Extract position metadata if present\nlet (old_path, new_path, old_line, new_line, position_type, lr_start, lr_end, base_sha, start_sha, head_sha) = note\n    .position\n    .as_ref()\n    .map(|pos| (\n        pos.old_path.clone(),\n        pos.new_path.clone(),\n        pos.old_line,\n        pos.new_line,\n        pos.position_type.clone(),                    // \"text\" | \"image\" | \"file\"\n        pos.line_range.as_ref().map(|r| r.start_line),\n        pos.line_range.as_ref().map(|r| r.end_line),\n        pos.base_sha.clone(),\n        pos.start_sha.clone(),\n        pos.head_sha.clone(),\n    ))\n    .unwrap_or((None, None, None, None, None, None, None, None, None, None));\n```\n\n## Strict Timestamp Parsing\n```rust\n// CRITICAL: Return error on invalid timestamps, never zero\nlet created_at = iso_to_ms(&note.created_at)\n    .ok_or_else(|| format\\!(\n        \"Invalid note.created_at for note {}: {}\",\n        note.id, note.created_at\n    ))?;\n```\n\n## NormalizedNote Fields for DiffNotes\n```rust\nNormalizedNote {\n    // ... existing fields ...\n    // DiffNote position metadata\n    position_old_path: old_path,\n    position_new_path: new_path,\n    position_old_line: old_line,\n    position_new_line: new_line,\n    // Extended position\n    position_type,\n    position_line_range_start: lr_start,\n    position_line_range_end: lr_end,\n    // SHA triplet\n    position_base_sha: base_sha,\n    position_start_sha: start_sha,\n    position_head_sha: head_sha,\n}\n```\n\n## Edge Cases\n- Notes without position should have all position fields as None\n- Invalid timestamp should fail the entire discussion (no partial results)\n- File renames: `old_path \\!= new_path` indicates a renamed file\n- Multi-line comments: `line_range` present means comment spans lines 45-48","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:41.208380Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:20:13.473091Z","closed_at":"2026-01-27T00:20:13.473031Z","close_reason":"Implemented transform_mr_discussion() and transform_notes_with_diff_position() with full DiffNote position extraction:\n- Extended NormalizedNote with 10 DiffNote position fields (path, line, type, line_range, SHA triplet)\n- Added strict timestamp parsing that returns Err on invalid timestamps\n- Created 13 diffnote_position_tests covering all extraction paths and error cases\n- Created 6 mr_discussion_tests verifying MR reference handling\n- All 161 tests passing","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-3j6","depends_on_id":"bd-3ir","type":"blocks","created_at":"2026-01-26T22:08:54.207801Z","created_by":"tayloreernisse"},{"issue_id":"bd-3j6","depends_on_id":"bd-5ta","type":"blocks","created_at":"2026-01-26T22:08:54.244201Z","created_by":"tayloreernisse"}]}
 {"id":"bd-3js","title":"Implement MR CLI commands (list, show, count)","description":"## Background\nCLI commands for viewing and filtering merge requests. Includes list, show, and count commands with MR-specific filters.\n\n## Approach\nUpdate existing CLI command files:\n1. `list.rs` - Add MR listing with filters\n2. `show.rs` - Add MR detail view with discussions\n3. `count.rs` - Add MR counting with state breakdown\n\n## Files\n- `src/cli/commands/list.rs` - Add MR subcommand\n- `src/cli/commands/show.rs` - Add MR detail view\n- `src/cli/commands/count.rs` - Add MR counting\n\n## Acceptance Criteria\n- [ ] `gi list mrs` shows MR table with iid, title, state, author, branches\n- [ ] `gi list mrs --state=merged` filters by state\n- [ ] `gi list mrs --state=locked` filters locally (not server-side)\n- [ ] `gi list mrs --draft` shows only draft MRs\n- [ ] `gi list mrs --no-draft` excludes draft MRs\n- [ ] `gi list mrs --reviewer=username` filters by reviewer\n- [ ] `gi list mrs --target-branch=main` filters by target branch\n- [ ] `gi list mrs --source-branch=feature/x` filters by source branch\n- [ ] Draft MRs show `[DRAFT]` prefix in title\n- [ ] `gi show mr <iid>` displays full detail including discussions\n- [ ] DiffNote shows file context: `[src/file.ts:45]`\n- [ ] Multi-line DiffNote shows: `[src/file.ts:45-48]`\n- [ ] `gi show mr` shows `detailed_merge_status`\n- [ ] `gi count mrs` shows total with state breakdown\n- [ ] `gi sync-status` shows MR cursor positions\n- [ ] `cargo test cli_commands` passes\n\n## TDD Loop\nRED: `cargo test list_mrs` -> command not found\nGREEN: Add MR subcommand\nVERIFY: `gi list mrs --help`\n\n## gi list mrs Output\n```\nMerge Requests (showing 20 of 1,234)\n\n  !847  Refactor auth to use JWT tokens           merged   @johndoe   main <- feature/jwt   3 days ago\n  !846  Fix memory leak in websocket handler      opened   @janedoe   main <- fix/websocket 5 days ago\n  !845  [DRAFT] Add dark mode CSS variables       opened   @bobsmith  main <- ui/dark-mode  1 week ago\n```\n\n## SQL for MR Listing\n```sql\nSELECT \n    m.iid, m.title, m.state, m.draft, m.author_username,\n    m.target_branch, m.source_branch, m.updated_at\nFROM merge_requests m\nWHERE m.project_id = ?\n  AND (? IS NULL OR m.state = ?)           -- state filter\n  AND (? IS NULL OR m.draft = ?)           -- draft filter\n  AND (? IS NULL OR m.author_username = ?) -- author filter\n  AND (? IS NULL OR m.target_branch = ?)   -- target-branch filter\n  AND (? IS NULL OR m.source_branch = ?)   -- source-branch filter\n  AND (? IS NULL OR EXISTS (               -- reviewer filter\n    SELECT 1 FROM mr_reviewers r \n    WHERE r.merge_request_id = m.id AND r.username = ?\n  ))\nORDER BY m.updated_at DESC\nLIMIT ?\n```\n\n## gi show mr Output\n```\nMerge Request !847: Refactor auth to use JWT tokens\n================================================================================\n\nProject:        group/project-one\nState:          merged\nDraft:          No\nAuthor:         @johndoe\nAssignees:      @janedoe, @bobsmith\nReviewers:      @alice, @charlie\nSource:         feature/jwt\nTarget:         main\nMerge Status:   mergeable\nMerged By:      @alice\nMerged At:      2024-03-20 14:30:00\nLabels:         enhancement, auth, reviewed\n\nDescription:\n  Moving away from session cookies to JWT-based authentication...\n\nDiscussions (8):\n\n  @janedoe (2024-03-16) [src/auth/jwt.ts:45]:\n    Should we use a separate signing key for refresh tokens?\n\n    @johndoe (2024-03-16):\n      Good point. I'll add a separate key with rotation support.\n\n  @alice (2024-03-18) [RESOLVED]:\n    Looks good! Just one nit about the token expiry constant.\n```\n\n## DiffNote File Context Display\n```rust\n// Build file context string\nlet file_context = match (note.position_new_path, note.position_new_line, note.position_line_range_end) {\n    (Some(path), Some(line), Some(end_line)) if line != end_line => {\n        format!(\"[{}:{}-{}]\", path, line, end_line)\n    }\n    (Some(path), Some(line), _) => {\n        format!(\"[{}:{}]\", path, line)\n    }\n    _ => String::new(),\n};\n```\n\n## gi count mrs Output\n```\nMerge Requests: 1,234\n  opened: 89\n  merged: 1,045\n  closed: 100\n```\n\n## Filter Arguments (clap)\n```rust\n#[derive(Parser)]\nstruct ListMrsArgs {\n    #[arg(long)]\n    state: Option<String>,  // opened|merged|closed|locked|all\n    #[arg(long)]\n    draft: bool,\n    #[arg(long)]\n    no_draft: bool,\n    #[arg(long)]\n    author: Option<String>,\n    #[arg(long)]\n    assignee: Option<String>,\n    #[arg(long)]\n    reviewer: Option<String>,\n    #[arg(long)]\n    target_branch: Option<String>,\n    #[arg(long)]\n    source_branch: Option<String>,\n    #[arg(long)]\n    label: Vec<String>,\n    #[arg(long)]\n    project: Option<String>,\n    #[arg(long, default_value = \"20\")]\n    limit: u32,\n}\n```\n\n## Edge Cases\n- `--state=locked` must filter locally (GitLab API doesn't support it)\n- Ambiguous MR iid across projects: prompt for `--project`\n- Empty discussions: show \"No discussions\" message\n- Multi-line DiffNotes: show line range in context","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:43.354939Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:37:31.792569Z","closed_at":"2026-01-27T00:37:31.792504Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-3js","depends_on_id":"bd-20h","type":"blocks","created_at":"2026-01-26T22:08:55.209249Z","created_by":"tayloreernisse"},{"issue_id":"bd-3js","depends_on_id":"bd-ser","type":"blocks","created_at":"2026-01-26T22:08:55.117728Z","created_by":"tayloreernisse"}]}
@@ -88,19 +112,27 @@
 {"id":"bd-3qm","title":"[CP1] Final validation - tests, smoke tests, integrity checks","description":"Run all tests and perform data integrity checks.\n\nValidation steps:\n1. Run all unit tests (vitest)\n2. Run all integration tests\n3. Run ESLint\n4. Run TypeScript strict check\n5. Manual smoke tests per PRD table\n6. Data integrity SQL checks:\n   - Issue count matches GitLab\n   - Every issue has raw_payload\n   - Labels in junction exist in labels table\n   - sync_cursors has entry per project\n   - Re-run fetches 0 new items\n   - Discussion count > 0\n   - Every discussion has >= 1 note\n   - individual_note=true has exactly 1 note\n\nFiles: All CP1 files\nDone when: All gate criteria from Definition of Done pass","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:20:51.994183Z","created_by":"tayloreernisse","updated_at":"2026-01-25T15:21:35.152852Z","deleted_at":"2026-01-25T15:21:35.152849Z","deleted_by":"tayloreernisse","delete_reason":"delete","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-3qs","title":"Implement lore generate-docs CLI command","description":"## Background\nThe generate-docs CLI command is the user-facing wrapper around the document regeneration pipeline. It has two modes: incremental (default, processes dirty_sources queue only) and full (seeds dirty_sources with ALL entities, then drains). Both modes use the same regenerator codepath to avoid logic divergence. Full mode uses keyset pagination (WHERE id > last_id) for seeding to avoid O(n^2) OFFSET degradation on large tables.\n\n## Approach\nCreate `src/cli/commands/generate_docs.rs` per PRD Section 2.4.\n\n**Core function:**\n```rust\npub fn run_generate_docs(\n    config: &Config,\n    full: bool,\n    project_filter: Option<&str>,\n) -> Result<GenerateDocsResult>\n```\n\n**Full mode seeding (keyset pagination):**\n```rust\nconst FULL_MODE_CHUNK_SIZE: usize = 2000;\n\n// For each source type (issues, MRs, discussions):\nlet mut last_id: i64 = 0;\nloop {\n    let tx = conn.transaction()?;\n    let inserted = tx.execute(\n        \"INSERT INTO dirty_sources (source_type, source_id, queued_at, ...)\n         SELECT 'issue', id, ?, 0, NULL, NULL, NULL\n         FROM issues WHERE id > ? ORDER BY id LIMIT ?\n         ON CONFLICT(source_type, source_id) DO NOTHING\",\n        params![now_ms(), last_id, FULL_MODE_CHUNK_SIZE],\n    )?;\n    if inserted == 0 { tx.commit()?; break; }\n    // Advance keyset cursor...\n    tx.commit()?;\n}\n```\n\n**After draining (full mode only):**\n```sql\nINSERT INTO documents_fts(documents_fts) VALUES('optimize')\n```\n\n**CLI args:**\n```rust\n#[derive(Args)]\npub struct GenerateDocsArgs {\n    #[arg(long)]\n    full: bool,\n    #[arg(long)]\n    project: Option<String>,\n}\n```\n\n**Output:** Human-readable table + JSON robot mode.\n\n## Acceptance Criteria\n- [ ] Default mode (no --full): processes only existing dirty_sources entries\n- [ ] --full mode: seeds dirty_sources with ALL issues, MRs, and discussions\n- [ ] Full mode uses keyset pagination (WHERE id > last_id, not OFFSET)\n- [ ] Full mode chunk size is 2000\n- [ ] Full mode does FTS optimize after completion\n- [ ] Both modes use regenerate_dirty_documents() (same codepath)\n- [ ] Progress bar shown in human mode (via indicatif)\n- [ ] JSON output in robot mode with GenerateDocsResult\n- [ ] GenerateDocsResult has issues/mrs/discussions/total/truncated/skipped counts\n- [ ] `cargo build` succeeds\n\n## Files\n- `src/cli/commands/generate_docs.rs` — new file\n- `src/cli/commands/mod.rs` — add `pub mod generate_docs;`\n- `src/cli/mod.rs` — add GenerateDocsArgs, wire up generate-docs subcommand\n- `src/main.rs` — add generate-docs command handler\n\n## TDD Loop\nRED: Integration test with seeded DB\nGREEN: Implement run_generate_docs with seeding + drain\nVERIFY: `cargo build && cargo test generate_docs`\n\n## Edge Cases\n- Empty database (no issues/MRs/discussions): full mode seeds nothing, returns all-zero counts\n- --project filter in full mode: only seed dirty_sources for entities in that project\n- Interrupted full mode: dirty_sources entries persist (ON CONFLICT DO NOTHING), resume by re-running\n- FTS optimize on empty FTS table: no-op (safe)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:25:55.226666Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:49:23.397157Z","closed_at":"2026-01-30T17:49:23.397098Z","close_reason":"Implemented generate-docs command with incremental + full mode, keyset pagination seeding, FTS optimize, project filter, human + JSON output. Builds clean.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-3qs","depends_on_id":"bd-1u1","type":"blocks","created_at":"2026-01-30T15:29:16.089769Z","created_by":"tayloreernisse"},{"issue_id":"bd-3qs","depends_on_id":"bd-221","type":"blocks","created_at":"2026-01-30T15:29:16.125158Z","created_by":"tayloreernisse"}]}
 {"id":"bd-3rl","title":"Epic: Gate C - Sync MVP","description":"## Background\nGate C adds the sync orchestrator and queue infrastructure that makes the search pipeline incremental and self-maintaining. It introduces dirty source tracking (change detection during ingestion), the discussion fetch queue, and the unified lore sync command that orchestrates the full pipeline. Gate C also adds integrity checks and repair paths.\n\n## Gate C Deliverables\n1. Orchestrated lore sync command with incremental doc regen + re-embedding\n2. Integrity checks + repair paths for FTS/embeddings consistency\n\n## Bead Dependencies (execution order, after Gate A)\n1. **bd-mem** — Shared backoff utility (no deps, shared with Gate B)\n2. **bd-38q** — Dirty source tracking (blocked by bd-36p, bd-hrs, bd-mem)\n3. **bd-1je** — Discussion queue (blocked by bd-hrs, bd-mem)\n4. **bd-1i2** — Integrate dirty tracking into ingestion (blocked by bd-38q)\n5. **bd-1x6** — Sync CLI (blocked by bd-38q, bd-1je, bd-1i2, bd-3qs, bd-2sx)\n\n## Acceptance Criteria\n- [ ] `lore sync` runs full pipeline: ingest -> generate-docs -> embed\n- [ ] `lore sync --full` does full re-sync + regeneration\n- [ ] `lore sync --no-embed` skips embedding stage\n- [ ] Dirty tracking: upserted entities automatically marked for regeneration\n- [ ] Queue draining: dirty_sources fully drained in bounded batch loop\n- [ ] Backoff: failed items use exponential backoff with jitter\n- [ ] `lore stats --check` detects inconsistencies\n- [ ] `lore stats --repair` fixes FTS/embedding inconsistencies","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-30T15:25:13.494698Z","created_by":"tayloreernisse","updated_at":"2026-01-30T18:05:52.121666Z","closed_at":"2026-01-30T18:05:52.121619Z","close_reason":"All Gate C sub-beads complete: backoff utility, dirty tracking, discussion queue, ingestion integration, sync CLI, stats CLI","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-3rl","depends_on_id":"bd-1x6","type":"blocks","created_at":"2026-01-30T15:29:35.853817Z","created_by":"tayloreernisse"},{"issue_id":"bd-3rl","depends_on_id":"bd-pr1","type":"blocks","created_at":"2026-01-30T15:29:35.892441Z","created_by":"tayloreernisse"}]}
+{"id":"bd-3sh","title":"Add 'lore count events' command with robot mode","description":"## Background\nNeed to verify event ingestion and report counts by type. The existing count command (src/cli/commands/count.rs) handles issues, mrs, discussions, notes with both human and robot output. This adds 'events' as a new count subcommand.\n\n## Approach\nExtend the existing count command in src/cli/commands/count.rs:\n\n1. Add CountTarget::Events variant (or string match) in the count dispatcher\n2. Query each event table with GROUP BY entity type:\n```sql\nSELECT \n  CASE WHEN issue_id IS NOT NULL THEN 'issue' ELSE 'merge_request' END as entity_type,\n  COUNT(*) as count\nFROM resource_state_events\nGROUP BY entity_type;\n-- (repeat for label and milestone events)\n```\n\n3. Human output: table format\n```\nEvent Type       Issues  MRs    Total\nState events     1,234   567    1,801\nLabel events     2,345   890    3,235\nMilestone events   456   123      579\nTotal            4,035   1,580  5,615\n```\n\n4. Robot JSON:\n```json\n{\n  \"ok\": true,\n  \"data\": {\n    \"state_events\": {\"issue\": 1234, \"merge_request\": 567, \"total\": 1801},\n    \"label_events\": {\"issue\": 2345, \"merge_request\": 890, \"total\": 3235},\n    \"milestone_events\": {\"issue\": 456, \"merge_request\": 123, \"total\": 579},\n    \"total\": 5615\n  }\n}\n```\n\n5. Register in CLI: add \"events\" to count's entity_type argument in src/cli/mod.rs\n\n## Acceptance Criteria\n- [ ] `lore count events` shows correct counts by event type and entity type\n- [ ] Robot JSON matches the schema above\n- [ ] Works with empty tables (all zeros)\n- [ ] Does not error if migration 011 hasn't been applied (graceful degradation or \"no event tables\" message)\n\n## Files\n- src/cli/commands/count.rs (add events counting logic)\n- src/cli/mod.rs (add \"events\" to count's accepted entity types)\n\n## TDD Loop\nRED: tests/count_tests.rs (or extend existing):\n- `test_count_events_empty_tables` - verify all zeros on fresh DB\n- `test_count_events_with_data` - seed state + label events, verify correct counts\n- `test_count_events_robot_json` - verify JSON structure\n\nGREEN: Add the events branch to count command\n\nVERIFY: `cargo test count -- --nocapture`\n\n## Edge Cases\n- Tables don't exist if user hasn't run migrate — check table existence first or catch the error\n- COUNT with GROUP BY returns no rows for empty tables — need to handle missing entity types as 0","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T21:31:57.379702Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:43:22.087169Z","compaction_level":0,"original_size":0,"labels":["cli","gate-1","phase-b"],"dependencies":[{"issue_id":"bd-3sh","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:57.380927Z","created_by":"tayloreernisse"},{"issue_id":"bd-3sh","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T21:32:06.308285Z","created_by":"tayloreernisse"}]}
 {"id":"bd-4qd","title":"Write unit tests for core algorithms","description":"## Background\nUnit tests verify the core algorithms in isolation: document extraction formatting, FTS query sanitization, RRF scoring, content hashing, backoff curves, and filter helpers. These tests don't require a database or external services — they test pure functions and logic.\n\n## Approach\nAdd #[cfg(test)] mod tests blocks to each module:\n\n**1. src/documents/extractor.rs:**\n- test_source_type_parse_all_aliases — every alias resolves correctly\n- test_source_type_parse_unknown — returns None\n- test_source_type_as_str_roundtrip — as_str matches parse input\n- test_content_hash_deterministic — same input = same hash\n- test_list_hash_order_independent — sorted before hashing\n- test_list_hash_empty — empty vec produces consistent hash\n\n**2. src/documents/truncation.rs:**\n- test_truncation_edge_cases (per bd-18t TDD Loop)\n\n**3. src/search/fts.rs:**\n- test_to_fts_query_basic — \"auth error\" -> quoted tokens\n- test_to_fts_query_prefix — \"auth*\" preserves prefix\n- test_to_fts_query_special_chars — \"C++\" quoted correctly\n- test_to_fts_query_dash — \"-DWITH_SSL\" quoted (not NOT operator)\n- test_to_fts_query_internal_quotes — escaped by doubling\n- test_to_fts_query_empty — empty string returns empty\n\n**4. src/search/rrf.rs:**\n- test_rrf_dual_list — docs in both lists score higher\n- test_rrf_normalization — best score = 1.0\n- test_rrf_empty — empty returns empty\n\n**5. src/core/backoff.rs:**\n- test_exponential_curve — delays double each attempt\n- test_cap_at_one_hour — high attempt_count capped\n- test_jitter_range — within [0.9, 1.1) factor\n\n**6. src/search/filters.rs:**\n- test_has_any_filter — true/false for various filter combos\n- test_clamp_limit — 0->20, 200->100, 50->50\n- test_path_filter_from_str — trailing slash = Prefix\n\n**7. src/search/hybrid.rs (hydration round-trip):**\n- test_single_round_trip_query — verify hydration SQL produces correct structure\n\n## Acceptance Criteria\n- [ ] All edge cases covered per PRD acceptance criteria\n- [ ] Tests are unit tests (no DB, no network, no Ollama)\n- [ ] `cargo test` passes with all new tests\n- [ ] No test depends on execution order\n- [ ] Tests cover: document extractor formats, truncation, RRF, hashing, FTS sanitization, backoff, filters\n\n## Files\n- In-module tests in: extractor.rs, truncation.rs, fts.rs, rrf.rs, backoff.rs, filters.rs, hybrid.rs\n\n## TDD Loop\nThese tests ARE the TDD loop for their respective beads. Each implementation bead should write its tests first (RED), then implement (GREEN).\nVERIFY: `cargo test`\n\n## Edge Cases\n- Tests with Unicode: include emoji, CJK characters in truncation tests\n- Tests with empty strings: empty queries, empty content, empty labels\n- Tests with boundary values: limit=0, limit=100, limit=101","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-30T15:27:21.712924Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:46:00.059346Z","closed_at":"2026-01-30T17:46:00.059292Z","close_reason":"All acceptance criteria tests already exist across modules. 276 tests passing (189 unit + 87 integration).","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-4qd","depends_on_id":"bd-18t","type":"blocks","created_at":"2026-01-30T15:29:35.356715Z","created_by":"tayloreernisse"},{"issue_id":"bd-4qd","depends_on_id":"bd-1k1","type":"blocks","created_at":"2026-01-30T15:29:35.320913Z","created_by":"tayloreernisse"},{"issue_id":"bd-4qd","depends_on_id":"bd-36p","type":"blocks","created_at":"2026-01-30T15:29:35.465589Z","created_by":"tayloreernisse"},{"issue_id":"bd-4qd","depends_on_id":"bd-3ez","type":"blocks","created_at":"2026-01-30T15:29:35.393455Z","created_by":"tayloreernisse"},{"issue_id":"bd-4qd","depends_on_id":"bd-mem","type":"blocks","created_at":"2026-01-30T15:29:35.427448Z","created_by":"tayloreernisse"}]}
 {"id":"bd-5ta","title":"Add GitLab MR types to types.rs","description":"## Background\nGitLab API types for merge requests. These structs define how we deserialize GitLab API responses. Must handle deprecated field aliases for backward compatibility with older GitLab instances.\n\n## Approach\nAdd new structs to `src/gitlab/types.rs`:\n- `GitLabMergeRequest` - Main MR struct with all fields\n- `GitLabReviewer` - Reviewer with optional approval state\n- `GitLabReferences` - Short and full reference strings\n\nUse serde `#[serde(alias = \"...\")]` for deprecated field fallbacks.\n\n## Files\n- `src/gitlab/types.rs` - Add new structs after existing GitLabIssue\n- `tests/fixtures/gitlab_merge_request.json` - Test fixture\n\n## Acceptance Criteria\n- [ ] `GitLabMergeRequest` struct exists with all fields from PRD\n- [ ] `detailed_merge_status` field exists (non-deprecated)\n- [ ] `#[serde(alias = \"merge_status\")]` on `merge_status_legacy` for fallback\n- [ ] `merge_user` field exists (non-deprecated)\n- [ ] `merged_by` field exists for fallback\n- [ ] `draft` and `work_in_progress` both exist (draft preferred, WIP fallback)\n- [ ] `sha` field maps to `head_sha` in transformer\n- [ ] `references: Option<GitLabReferences>` for short/full refs\n- [ ] `state: String` supports \"opened\", \"merged\", \"closed\", \"locked\"\n- [ ] Fixture deserializes without error\n- [ ] `cargo test` passes\n\n## TDD Loop\nRED: Add test that deserializes fixture -> struct not found\nGREEN: Add GitLabMergeRequest, GitLabReviewer, GitLabReferences structs\nVERIFY: `cargo test gitlab_types`\n\n## Struct Definitions (from PRD)\n```rust\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabMergeRequest {\n    pub id: i64,\n    pub iid: i64,\n    pub project_id: i64,\n    pub title: String,\n    pub description: Option<String>,\n    pub state: String,                // \"opened\" | \"merged\" | \"closed\" | \"locked\"\n    #[serde(default)]\n    pub draft: bool,\n    #[serde(default)]\n    pub work_in_progress: bool,       // Deprecated fallback\n    pub source_branch: String,\n    pub target_branch: String,\n    pub sha: Option<String>,          // head_sha\n    pub references: Option<GitLabReferences>,\n    pub detailed_merge_status: Option<String>,\n    #[serde(alias = \"merge_status\")]\n    pub merge_status_legacy: Option<String>,\n    pub created_at: String,\n    pub updated_at: String,\n    pub merged_at: Option<String>,\n    pub closed_at: Option<String>,\n    pub author: GitLabAuthor,\n    pub merge_user: Option<GitLabAuthor>,\n    pub merged_by: Option<GitLabAuthor>,\n    #[serde(default)]\n    pub labels: Vec<String>,\n    #[serde(default)]\n    pub assignees: Vec<GitLabAuthor>,\n    #[serde(default)]\n    pub reviewers: Vec<GitLabReviewer>,\n    pub web_url: String,\n}\n\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabReferences {\n    pub short: String,   // e.g. \"\\!123\"\n    pub full: String,    // e.g. \"group/project\\!123\"\n}\n\n#[derive(Debug, Clone, Deserialize)]\npub struct GitLabReviewer {\n    pub id: i64,\n    pub username: String,\n    pub name: String,\n}\n```\n\n## Test Fixture (create tests/fixtures/gitlab_merge_request.json)\n```json\n{\n  \"id\": 12345,\n  \"iid\": 42,\n  \"project_id\": 100,\n  \"title\": \"Add user authentication\",\n  \"description\": \"Implements JWT auth flow\",\n  \"state\": \"merged\",\n  \"draft\": false,\n  \"work_in_progress\": false,\n  \"source_branch\": \"feature/auth\",\n  \"target_branch\": \"main\",\n  \"sha\": \"abc123def456\",\n  \"references\": { \"short\": \"\\!42\", \"full\": \"group/project\\!42\" },\n  \"detailed_merge_status\": \"mergeable\",\n  \"merge_status\": \"can_be_merged\",\n  \"created_at\": \"2024-01-15T10:00:00Z\",\n  \"updated_at\": \"2024-01-20T14:30:00Z\",\n  \"merged_at\": \"2024-01-20T14:30:00Z\",\n  \"closed_at\": null,\n  \"author\": { \"id\": 1, \"username\": \"johndoe\", \"name\": \"John Doe\" },\n  \"merge_user\": { \"id\": 2, \"username\": \"janedoe\", \"name\": \"Jane Doe\" },\n  \"merged_by\": { \"id\": 2, \"username\": \"janedoe\", \"name\": \"Jane Doe\" },\n  \"labels\": [\"enhancement\", \"auth\"],\n  \"assignees\": [{ \"id\": 3, \"username\": \"bob\", \"name\": \"Bob Smith\" }],\n  \"reviewers\": [{ \"id\": 4, \"username\": \"alice\", \"name\": \"Alice Wong\" }],\n  \"web_url\": \"https://gitlab.example.com/group/project/-/merge_requests/42\"\n}\n```\n\n## Edge Cases\n- `locked` state is transitional (merge in progress) - rare but valid\n- Some older instances may not return `detailed_merge_status`\n- Some older instances may not return `merge_user` (use `merged_by` fallback)\n- `work_in_progress` is deprecated but still returned by some instances","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:40.498088Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:08:35.520229Z","closed_at":"2026-01-27T00:08:35.520167Z","close_reason":"Added GitLabMergeRequest, GitLabReviewer, GitLabReferences structs. Updated GitLabNotePosition with position_type, line_range, and SHA triplet fields. All 23 type tests passing.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-5ta","depends_on_id":"bd-3ir","type":"blocks","created_at":"2026-01-26T22:08:53.981911Z","created_by":"tayloreernisse"}]}
 {"id":"bd-88m","title":"[CP1] Issue ingestion module","description":"Fetch and store issues with cursor-based incremental sync.\n\n## Module\nsrc/ingestion/issues.rs\n\n## Key Structs\n\n### IngestIssuesResult\n- fetched: usize\n- upserted: usize\n- labels_created: usize\n- issues_needing_discussion_sync: Vec<IssueForDiscussionSync>\n\n### IssueForDiscussionSync\n- local_issue_id: i64\n- iid: i64\n- updated_at: i64\n\n## Main Function\npub async fn ingest_issues(conn, client, config, project_id, gitlab_project_id) -> Result<IngestIssuesResult>\n\n## Logic\n1. Get current cursor from sync_cursors (updated_at_cursor, tie_breaker_id)\n2. Paginate through issues updated after cursor with cursor_rewind_seconds\n3. Apply local filtering for tuple cursor semantics:\n   - Skip if issue.updated_at < cursor_updated_at\n   - Skip if issue.updated_at == cursor_updated_at AND issue.id <= cursor_gitlab_id\n4. For each issue passing filter:\n   - Begin transaction\n   - Store raw payload (compressed)\n   - Transform and upsert issue\n   - Clear existing label links (DELETE FROM issue_labels)\n   - Extract and upsert labels\n   - Link issue to labels via junction\n   - Commit transaction\n   - Track for discussion sync eligibility\n5. Incremental cursor update every 100 issues\n6. Final cursor update\n7. Determine issues needing discussion sync: where updated_at > discussions_synced_for_updated_at\n\n## Helper Functions\n- get_cursor(conn, project_id) -> (Option<i64>, Option<i64>)\n- get_discussions_synced_at(conn, issue_id) -> Option<i64>\n- upsert_issue(conn, issue, payload_id) -> usize\n- get_local_issue_id(conn, gitlab_id) -> i64\n- clear_issue_labels(conn, issue_id)\n- upsert_label(conn, label) -> bool\n- get_label_id(conn, project_id, name) -> i64\n- link_issue_label(conn, issue_id, label_id)\n- update_cursor(conn, project_id, resource_type, updated_at, gitlab_id)\n\nFiles: src/ingestion/mod.rs, src/ingestion/issues.rs\nTests: tests/issue_ingestion_tests.rs\nDone when: Issues, labels, issue_labels populated correctly with resumable cursor","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T16:57:35.655708Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.806982Z","deleted_at":"2026-01-25T17:02:01.806977Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-8t4","title":"Extract cross-references from resource_state_events","description":"## Background\nresource_state_events includes source_merge_request (with iid) for 'closed by MR' events. After state events are stored (Gate 1), post-processing extracts these into entity_references for the cross-reference graph.\n\n## Approach\nCreate src/core/references.rs (new module) or add to events_db.rs:\n\n```rust\n/// Extract cross-references from stored state events and insert into entity_references.\n/// Looks for state events with source_merge_request_id IS NOT NULL (meaning \"closed by MR\").\n/// \n/// Directionality: source = MR (that caused the close), target = issue (that was closed)\npub fn extract_refs_from_state_events(\n    conn: &Connection,\n    project_id: i64,\n) -> Result<usize>  // returns count of new references inserted\n```\n\nSQL logic:\n```sql\nINSERT OR IGNORE INTO entity_references (\n    source_entity_type, source_entity_id,\n    target_entity_type, target_entity_id,\n    reference_type, source_method, created_at\n)\nSELECT\n    'merge_request',\n    mr.id,\n    'issue',\n    rse.issue_id,\n    'closes',\n    'api_state_event',\n    rse.created_at\nFROM resource_state_events rse\nJOIN merge_requests mr ON mr.project_id = rse.project_id AND mr.iid = rse.source_merge_request_id\nWHERE rse.source_merge_request_id IS NOT NULL\n  AND rse.issue_id IS NOT NULL\n  AND rse.project_id = ?1;\n```\n\nKey: source_merge_request_id stores the MR iid, so we JOIN on merge_requests.iid to get the local DB id.\n\nRegister in src/core/mod.rs: `pub mod references;`\n\nCall this after drain_dependent_queue in the sync pipeline (after all state events are stored).\n\n## Acceptance Criteria\n- [ ] State events with source_merge_request_id produce 'closes' references\n- [ ] Source = MR (resolved by iid), target = issue\n- [ ] source_method = 'api_state_event'\n- [ ] INSERT OR IGNORE prevents duplicates with api_closes_issues data\n- [ ] Returns count of newly inserted references\n- [ ] No-op when no state events have source_merge_request_id\n\n## Files\n- src/core/references.rs (new)\n- src/core/mod.rs (add `pub mod references;`)\n- src/cli/commands/sync.rs (call after drain step)\n\n## TDD Loop\nRED: tests/references_tests.rs:\n- `test_extract_refs_from_state_events_basic` - seed a \"closed\" state event with source_merge_request_id, verify entity_reference created\n- `test_extract_refs_dedup_with_closes_issues` - insert ref from closes_issues API first, verify state event extraction doesn't duplicate\n- `test_extract_refs_no_source_mr` - state events without source_merge_request_id produce no refs\n\nSetup: create_test_db with migrations 001-011, seed project + issue + MR + state events.\n\nGREEN: Implement extract_refs_from_state_events\n\nVERIFY: `cargo test references -- --nocapture`\n\n## Edge Cases\n- source_merge_request_id may reference an MR not synced locally (cross-project close) — the JOIN will produce no match, which is correct behavior (ref simply not created)\n- Multiple state events can reference the same MR for the same issue (reopen + re-close) — INSERT OR IGNORE handles dedup\n- The merge_requests table might not have the MR yet if sync is still running — call this after all dependent fetches complete","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:32:33.619606Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:41:50.562956Z","compaction_level":0,"original_size":0,"labels":["extraction","gate-2","phase-b"],"dependencies":[{"issue_id":"bd-8t4","depends_on_id":"bd-1ep","type":"blocks","created_at":"2026-02-02T21:32:42.945176Z","created_by":"tayloreernisse"},{"issue_id":"bd-8t4","depends_on_id":"bd-1se","type":"parent-child","created_at":"2026-02-02T21:32:33.621025Z","created_by":"tayloreernisse"},{"issue_id":"bd-8t4","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T22:41:50.562935Z","created_by":"tayloreernisse"}]}
 {"id":"bd-9av","title":"[CP1] gi sync-status enhancement","description":"Enhance sync-status from CP0 stub to show issue cursors.\n\n## Changes to src/cli/commands/sync_status.rs\n\nUpdate the existing stub to show:\n- Last run timestamp and duration\n- Cursor positions per project (issues resource_type)\n- Entity counts (issues, discussions, notes)\n\n## Output Format\nLast sync: 2026-01-25 10:30:00 (succeeded, 45s)\n\nCursors:\n  group/project-one\n    issues: 2026-01-25T10:25:00Z (gitlab_id: 12345678)\n\nCounts:\n  Issues: 1,234\n  Discussions: 5,678\n  Notes: 23,456 (4,567 system)\n\nFiles: src/cli/commands/sync_status.rs\nDone when: Shows cursor positions and counts after ingestion","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:58:27.246825Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.968507Z","deleted_at":"2026-01-25T17:02:01.968503Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-9dd","title":"Implement 'lore trace' command with human and robot output","description":"## Background\nlore trace <file> shows the MR→issue→discussion chain for a file. CLI wiring with human and robot output per spec §5.5.\n\n## Approach\nCreate src/cli/commands/trace.rs:\n\n**1. CLI args (in src/cli/mod.rs):**\n```rust\n/// Trace a file to its motivating decisions\n#[command(name = \"trace\")]\nTrace(TraceArgs),\n\n#[derive(clap::Args)]\npub struct TraceArgs {\n    /// File path to trace (optionally :line for future Tier 2)\n    #[arg(required = true)]\n    pub path: String,\n\n    /// Scope to a specific project\n    #[arg(short = 'p', long = \"project\")]\n    pub project: Option<String>,\n\n    /// Disable rename chain resolution\n    #[arg(long = \"no-follow-renames\")]\n    pub no_follow_renames: bool,\n\n    /// Maximum number of MRs to show\n    #[arg(short = 'n', long = \"limit\", default_value = \"20\")]\n    pub limit: usize,\n}\n```\n\n**2. Path parsing:**\n```rust\nfn parse_trace_path(input: &str) -> (String, Option<u32>) {\n    // \"src/foo.rs:45\" → (\"src/foo.rs\", Some(45))\n    // \"src/foo.rs\" → (\"src/foo.rs\", None)\n    if let Some((path, line)) = input.rsplit_once(':') {\n        if let Ok(n) = line.parse::<u32>() {\n            return (path.to_string(), Some(n));\n        }\n    }\n    (input.to_string(), None)\n}\n```\n\nIf line number provided, warn: \"Line-level tracing requires git integration (future feature). Proceeding with file-level trace.\"\n\n**3. Human output** (spec §5.5):\n```\nTrace: src/auth/oauth.rs\n────────────────────────\n\n!567  feat: add OAuth2 provider                    MERGED 2024-03-25\n  → Closes #234: Migrate to OAuth2\n  → 12 discussion comments, 4 on this file\n  → Decision: Use rust-oauth2 crate (discussed in #234, comment by @alice)\n\n!612  fix: token refresh race condition             MERGED 2024-04-10\n  → Closes #299: OAuth2 login fails for SSO users\n  → 5 discussion comments, 2 on this file\n  → [src/auth/oauth.rs:45] \"Add mutex around refresh to prevent double-refresh\"\n```\n\n**4. Robot JSON:**\n```json\n{\n  \"ok\": true,\n  \"data\": {\n    \"file_path\": \"src/auth/oauth.rs\",\n    \"rename_chain\": [\"src/auth/handler.rs\", \"src/auth/oauth.rs\"],\n    \"trace_chains\": [{\n      \"mr\": { \"iid\": 567, \"title\": \"...\", \"state\": \"merged\", ... },\n      \"issues\": [{ \"iid\": 234, \"title\": \"...\", \"reference_type\": \"closes\", ... }],\n      \"discussions\": [{ \"note_id\": 123, \"snippet\": \"...\", \"position_line\": 45, ... }]\n    }]\n  },\n  \"meta\": { \"total_mrs\": 2, \"tier\": \"api_only\" }\n}\n```\n\n**5. Graceful empty state:**\n\"No MR data found for this file. Run lore sync with fetchMrFileChanges: true\"\n\n## Acceptance Criteria\n- [ ] `lore trace src/auth/oauth.rs` shows trace chains\n- [ ] Human output matches spec §5.5 format\n- [ ] Robot JSON structured with trace_chains array\n- [ ] :line suffix parsed with Tier 2 warning\n- [ ] -p flag for project scoping\n- [ ] --no-follow-renames disables rename chain\n- [ ] Graceful empty state message\n- [ ] meta.tier = \"api_only\" (Tier 1)\n\n## Files\n- src/cli/commands/trace.rs (new)\n- src/cli/commands/mod.rs (add `pub mod trace;`)\n- src/cli/mod.rs (add Trace variant + TraceArgs)\n- src/main.rs (add handler)\n\n## TDD Loop\nRED: tests/trace_command_tests.rs:\n- `test_trace_command_basic` - end-to-end with seeded data\n- `test_trace_command_empty_state` - no data, verify message\n- `test_trace_command_line_number_warning` - :45 suffix, verify warning\n- `test_trace_command_robot_json` - verify JSON structure\n- `test_parse_trace_path` - unit test path parsing\n\nGREEN: Implement CLI wiring + output renderers\n\nVERIFY: `cargo test trace_command -- --nocapture && cargo build`\n\n## Edge Cases\n- Path with colons in directory name (unlikely but possible on macOS): rsplit_once handles this (last colon)\n- File path that doesn't exist in mr_file_changes: empty state, not error\n- Very long DiffNote snippets: truncate to ~200 chars","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:34:32.788530Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:13.552923Z","compaction_level":0,"original_size":0,"labels":["cli","gate-5","phase-b"],"dependencies":[{"issue_id":"bd-9dd","depends_on_id":"bd-1ht","type":"parent-child","created_at":"2026-02-02T21:34:32.789920Z","created_by":"tayloreernisse"},{"issue_id":"bd-9dd","depends_on_id":"bd-2n4","type":"blocks","created_at":"2026-02-02T21:34:37.941327Z","created_by":"tayloreernisse"}]}
 {"id":"bd-am7","title":"Implement embedding pipeline with chunking","description":"## Background\nThe embedding pipeline takes documents, chunks them (paragraph-boundary splitting with overlap), sends chunks to Ollama for embedding via async HTTP, and stores vectors in sqlite-vec + metadata. It uses keyset pagination, concurrent HTTP requests via FuturesUnordered, per-batch transactions, and dimension validation.\n\n## Approach\nCreate \\`src/embedding/pipeline.rs\\` per PRD Section 4.4. **The pipeline is async.**\n\n**Constants (per PRD):**\n```rust\nconst BATCH_SIZE: usize = 32;           // texts per Ollama API call\nconst DB_PAGE_SIZE: usize = 500;        // keyset pagination page size\nconst EXPECTED_DIMS: usize = 768;       // nomic-embed-text dimensions\nconst CHUNK_MAX_CHARS: usize = 32_000;  // max chars per chunk\nconst CHUNK_OVERLAP_CHARS: usize = 500; // overlap between chunks\n```\n\n**Core async function:**\n```rust\npub async fn embed_documents(\n    conn: &Connection,\n    client: &OllamaClient,\n    selection: EmbedSelection,\n    concurrency: usize,              // max in-flight HTTP requests\n    progress_callback: Option<Box<dyn Fn(usize, usize)>>,\n) -> Result<EmbedResult>\n```\n\n**EmbedSelection:** Pending | RetryFailed\n**EmbedResult:** { embedded, failed, skipped }\n\n**Algorithm (per PRD):**\n1. count_pending_documents(conn, selection) for progress total\n2. Keyset pagination loop: find_pending_documents(conn, DB_PAGE_SIZE, last_id, selection)\n3. For each page:\n   a. Begin transaction\n   b. For each doc: clear_document_embeddings(&tx, doc.id), split_into_chunks(&doc.content)\n   c. Build ChunkWork items with doc_hash + chunk_hash\n   d. Commit clearing transaction\n4. Batch ChunkWork texts into Ollama calls (BATCH_SIZE=32)\n5. Use **FuturesUnordered** for concurrent HTTP, cap at \\`concurrency\\`\n6. collect_writes() in per-batch transactions: validate dims (768), store LE bytes, write metadata\n7. On error: record_embedding_error per chunk (not abort)\n8. Advance keyset cursor\n\n**ChunkWork struct:**\n```rust\nstruct ChunkWork {\n    doc_id: i64,\n    chunk_index: usize,\n    doc_hash: String,    // SHA-256 of FULL document (staleness detection)\n    chunk_hash: String,  // SHA-256 of THIS chunk (provenance)\n    text: String,\n}\n```\n\n**Splitting:** split_into_chunks(content) -> Vec<(usize, String)>\n- Documents <= CHUNK_MAX_CHARS: single chunk (index 0)\n- Longer: split at paragraph boundaries (\\\\n\\\\n), fallback to sentence/word, with CHUNK_OVERLAP_CHARS overlap\n\n**Storage:** embeddings as raw LE bytes, rowid = encode_rowid(doc_id, chunk_idx)\n**Staleness detection:** uses document_hash (not chunk_hash) because it's document-level\n\nAlso create \\`src/embedding/change_detector.rs\\` (referenced in PRD module structure):\n```rust\npub fn detect_embedding_changes(conn: &Connection) -> Result<Vec<PendingDocument>>;\n```\n\n## Acceptance Criteria\n- [ ] Pipeline is async (uses FuturesUnordered for concurrent HTTP)\n- [ ] concurrency parameter caps in-flight HTTP requests\n- [ ] progress_callback reports (processed, total)\n- [ ] New documents embedded, changed re-embedded, unchanged skipped\n- [ ] clear_document_embeddings before re-embedding (range delete vec0 + metadata)\n- [ ] Chunking at paragraph boundaries with 500-char overlap\n- [ ] Short documents (<32k chars) produce exactly 1 chunk\n- [ ] Embeddings stored as raw LE bytes in vec0\n- [ ] Rowids encoded via encode_rowid(doc_id, chunk_index)\n- [ ] Dimension validation: 768 floats per embedding (mismatch -> record error, not store)\n- [ ] Per-batch transactions for writes\n- [ ] Errors recorded in embedding_metadata per chunk (last_error, attempt_count)\n- [ ] Keyset pagination (d.id > last_id, not OFFSET)\n- [ ] Pending detection uses document_hash (not chunk_hash)\n- [ ] \\`cargo build\\` succeeds\n\n## Files\n- \\`src/embedding/pipeline.rs\\` — new file (async)\n- \\`src/embedding/change_detector.rs\\` — new file\n- \\`src/embedding/mod.rs\\` — add \\`pub mod pipeline; pub mod change_detector;\\` + re-exports\n\n## TDD Loop\nRED: Unit tests for chunking:\n- \\`test_short_document_single_chunk\\` — <32k produces [(0, full_content)]\n- \\`test_long_document_multiple_chunks\\` — >32k splits at paragraph boundaries\n- \\`test_chunk_overlap\\` — adjacent chunks share 500-char overlap\n- \\`test_no_paragraph_boundary\\` — falls back to char boundary\nIntegration tests need Ollama or mock.\nGREEN: Implement split_into_chunks, embed_documents (async)\nVERIFY: \\`cargo test pipeline\\`\n\n## Edge Cases\n- Empty document content_text: skip (don't embed)\n- No paragraph boundaries: split at CHUNK_MAX_CHARS with overlap\n- Ollama error for one batch: record error per chunk, continue with next batch\n- Dimension mismatch (model returns 512 instead of 768): record error, don't store corrupt data\n- Document deleted between pagination and embedding: skip gracefully","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:34.093701Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:58:58.908585Z","closed_at":"2026-01-30T17:58:58.908525Z","close_reason":"Implemented embedding pipeline: chunking at paragraph boundaries with 500-char overlap, change detector (keyset pagination, hash-based staleness), async embed via Ollama with batch processing, dimension validation, per-chunk error recording, LE byte vector storage. 7 chunking tests pass. 289 total tests.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-am7","depends_on_id":"bd-1y8","type":"blocks","created_at":"2026-01-30T15:29:24.697418Z","created_by":"tayloreernisse"},{"issue_id":"bd-am7","depends_on_id":"bd-2ac","type":"blocks","created_at":"2026-01-30T15:29:24.732567Z","created_by":"tayloreernisse"},{"issue_id":"bd-am7","depends_on_id":"bd-335","type":"blocks","created_at":"2026-01-30T15:29:24.660199Z","created_by":"tayloreernisse"}]}
 {"id":"bd-bjo","title":"Implement vector search function","description":"## Background\nVector search queries the sqlite-vec virtual table for nearest-neighbor documents. Because documents may have multiple chunks, the raw KNN results need deduplication by document_id (keeping the best/lowest distance per document). The function over-fetches 3x to ensure enough unique documents after dedup.\n\n## Approach\nCreate `src/search/vector.rs`:\n\n```rust\npub struct VectorResult {\n    pub document_id: i64,\n    pub distance: f64,  // Lower = closer match\n}\n\n/// Search documents using sqlite-vec KNN query.\n/// Over-fetches 3x limit to handle chunk dedup.\npub fn search_vector(\n    conn: &Connection,\n    query_embedding: &[f32],  // 768-dim embedding of search query\n    limit: usize,\n) -> Result<Vec<VectorResult>>\n```\n\n**SQL (KNN query):**\n```sql\nSELECT rowid, distance\nFROM embeddings\nWHERE embedding MATCH ?\n  AND k = ?\nORDER BY distance\n```\n\n**Algorithm:**\n1. Convert query_embedding to raw LE bytes\n2. Execute KNN with k = limit * 3 (over-fetch for dedup)\n3. Decode each rowid via decode_rowid() -> (document_id, chunk_index)\n4. Group by document_id, keep minimum distance (best chunk)\n5. Sort by distance ascending\n6. Take first `limit` results\n\n## Acceptance Criteria\n- [ ] Returns deduplicated document-level results (not chunk-level)\n- [ ] Best chunk distance kept per document (lowest distance wins)\n- [ ] KNN with k parameter (3x limit)\n- [ ] Query embedding passed as raw LE bytes\n- [ ] Results sorted by distance ascending (closest first)\n- [ ] Returns at most `limit` results\n- [ ] Empty embeddings table returns empty Vec\n- [ ] `cargo build` succeeds\n\n## Files\n- `src/search/vector.rs` — new file\n- `src/search/mod.rs` — add `pub use vector::{search_vector, VectorResult};`\n\n## TDD Loop\nRED: Integration tests need sqlite-vec + seeded embeddings:\n- `test_vector_search_basic` — finds nearest document\n- `test_vector_search_dedup` — multi-chunk doc returns once with best distance\n- `test_vector_search_empty` — empty table returns empty\n- `test_vector_search_limit` — respects limit parameter\nGREEN: Implement search_vector\nVERIFY: `cargo test vector`\n\n## Edge Cases\n- All chunks belong to same document: returns single result\n- Query embedding wrong dimension: sqlite-vec may error — handle gracefully\n- Over-fetch returns fewer than limit unique docs: return what we have\n- Distance = 0.0: exact match (valid result)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:50.270357Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:44:56.233611Z","closed_at":"2026-01-30T17:44:56.233512Z","close_reason":"Implemented search_vector with KNN query, 3x over-fetch, chunk dedup. 3 tests pass.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-bjo","depends_on_id":"bd-1y8","type":"blocks","created_at":"2026-01-30T15:29:24.842469Z","created_by":"tayloreernisse"},{"issue_id":"bd-bjo","depends_on_id":"bd-2ac","type":"blocks","created_at":"2026-01-30T15:29:24.878048Z","created_by":"tayloreernisse"}]}
 {"id":"bd-cbo","title":"[CP1] Cargo.toml updates - async-stream and futures","description":"Add required dependencies for async pagination streams.\n\n## Changes\nAdd to Cargo.toml:\n- async-stream = \"0.3\"\n- futures = \"0.3\"\n\n## Why\nThe pagination methods use async generators which require async-stream crate.\nfutures crate provides StreamExt for consuming the streams.\n\n## Done When\n- cargo check passes with new deps\n- No unused dependency warnings\n\nFiles: Cargo.toml","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:42:31.143927Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.661666Z","deleted_at":"2026-01-25T17:02:01.661662Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-cq2","title":"[CP1] Integration tests for label linkage","description":"Integration tests verifying label linkage and stale removal.\n\n## Tests (tests/label_linkage_tests.rs)\n\n- clears_existing_labels_before_linking_new_set\n- removes_stale_label_links_on_issue_update\n- handles_issue_with_all_labels_removed\n- preserves_labels_that_still_exist\n\n## Test Scenario\n1. Create issue with labels [A, B]\n2. Verify issue_labels has links to A and B\n3. Update issue with labels [B, C]\n4. Verify A link removed, B preserved, C added\n\n## Why This Matters\nThe clear-and-relink pattern ensures GitLab reality is reflected locally.\nIf we only INSERT, removed labels would persist incorrectly.\n\nFiles: tests/label_linkage_tests.rs\nDone when: Stale label links correctly removed on resync","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:59:10.665771Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:02.062192Z","deleted_at":"2026-01-25T17:02:02.062188Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-czk","title":"Add entity_references table to migration 010","description":"## Background\nThe entity_references table is now part of migration 011 (combined with resource event tables and dependent fetch queue). This bead is satisfied by bd-hu3 since the entity_references table schema is included in the same migration.\n\n## Approach\nThis bead's work is folded into bd-hu3 (Write migration 011). The entity_references table from Phase B spec §2.2 is included in migrations/011_resource_events.sql alongside the event tables and queue.\n\nThe entity_references schema includes:\n- source/target entity type + id with reference_type and source_method\n- Unresolved reference support (target_entity_id NULL with target_project_path + target_entity_iid)\n- UNIQUE constraint using COALESCE for nullable columns\n- Partial indexes for source, target (where not null), and unresolved refs\n\nNo separate migration file needed — this is in 011.\n\n## Acceptance Criteria\n- [ ] entity_references table exists in migration 011 (verified by bd-hu3)\n- [ ] UNIQUE constraint handles NULL columns via COALESCE\n- [ ] Indexes created: source composite, target composite (partial), unresolved (partial)\n- [ ] reference_type CHECK includes 'closes', 'mentioned', 'related'\n- [ ] source_method CHECK includes 'api_closes_issues', 'api_state_event', 'system_note_parse'\n\n## Files\n- migrations/011_resource_events.sql (part of bd-hu3)\n\n## TDD Loop\nCovered by bd-hu3's test_migration_011_entity_references_dedup test.\n\nVERIFY: `cargo test migration_tests -- --nocapture`\n\n## Edge Cases\n- Same as bd-hu3's entity_references edge cases","status":"closed","priority":2,"issue_type":"task","created_at":"2026-02-02T21:32:33.506883Z","created_by":"tayloreernisse","updated_at":"2026-02-02T22:42:06.104237Z","closed_at":"2026-02-02T22:42:06.104190Z","close_reason":"Work folded into bd-hu3 (migration 011 includes entity_references table)","compaction_level":0,"original_size":0,"labels":["gate-2","phase-b","schema"]}
+{"id":"bd-dty","title":"Implement timeline robot mode JSON output","description":"## Background\nRobot mode JSON for timeline command per spec §3.5. Must follow the existing robot mode pattern: {ok: true, data: {...}, meta: {...}}.\n\n## Approach\nAdd to src/cli/commands/timeline.rs:\n\n```rust\nuse serde::Serialize;\n\n#[derive(Serialize)]\nstruct TimelineOutput {\n    ok: bool,\n    data: TimelineDataJson,\n    meta: TimelineMetaJson,\n}\n\n#[derive(Serialize)]\nstruct TimelineDataJson {\n    query: String,\n    event_count: usize,\n    seed_entities: Vec<EntityRefJson>,\n    expanded_entities: Vec<ExpandedEntityRefJson>,\n    unresolved_references: Vec<UnresolvedRefJson>,\n    events: Vec<TimelineEventJson>,\n}\n\n#[derive(Serialize)]\nstruct TimelineEventJson {\n    timestamp: String,              // ISO 8601\n    entity_type: String,\n    entity_iid: i64,\n    project: String,\n    event_type: String,             // \"created\", \"state_changed\", etc.\n    summary: String,\n    actor: Option<String>,\n    url: Option<String>,\n    is_seed: bool,\n    details: serde_json::Value,     // type-specific details\n}\n\n#[derive(Serialize)]\nstruct TimelineMetaJson {\n    search_mode: String,            // \"lexical\"\n    expansion_depth: usize,\n    expand_mentions: bool,\n    total_entities: usize,\n    total_events: usize,\n    evidence_notes_included: usize,\n    unresolved_references: usize,\n}\n\npub fn print_timeline_json(result: &TimelineResult, depth: usize, expand_mentions: bool) {\n    let output = TimelineOutput {\n        ok: true,\n        data: TimelineDataJson {\n            query: result.query.clone(),\n            event_count: result.events.len(),\n            seed_entities: result.seed_entities.iter().map(|e| EntityRefJson::from(e)).collect(),\n            expanded_entities: result.expanded_entities.iter().map(|e| ExpandedEntityRefJson::from(e)).collect(),\n            unresolved_references: result.unresolved_references.iter().map(|r| UnresolvedRefJson::from(r)).collect(),\n            events: result.events.iter().map(|e| TimelineEventJson::from(e)).collect(),\n        },\n        meta: TimelineMetaJson { ... },\n    };\n    println!(\"{}\", serde_json::to_string(&output).unwrap());\n}\n```\n\n**Timestamp conversion:** ms epoch → ISO 8601 string (e.g., \"2024-03-15T10:00:00Z\")\n\n**Event type details mapping:**\n- Created → { \"labels\": [...] } (if entity has labels at creation)\n- StateChanged → { \"state\": \"closed\" }\n- LabelAdded → { \"label\": \"auth\" }\n- NoteEvidence → { \"note_id\": 12345, \"snippet\": \"...\" }\n- Merged → {} (no extra details)\n\n## Acceptance Criteria\n- [ ] JSON matches spec §3.5 exactly\n- [ ] Timestamps in ISO 8601\n- [ ] All fields populated (seed_entities, expanded_entities with depth+via, events)\n- [ ] meta section includes query parameters (depth, expand_mentions)\n- [ ] Unresolved references included in JSON output\n- [ ] evidence_notes_included count in meta\n\n## Files\n- src/cli/commands/timeline.rs (add robot output function)\n\n## TDD Loop\nRED: tests/timeline_json_tests.rs:\n- `test_timeline_json_structure` - parse output as serde_json::Value, verify top-level keys\n- `test_timeline_json_event_timestamp_iso` - verify timestamps are ISO 8601\n- `test_timeline_json_expanded_provenance` - verify via field in expanded_entities\n- `test_timeline_json_unresolved_refs` - verify unresolved_references array\n- `test_timeline_json_meta_fields` - verify meta contains all required fields\n\nGREEN: Implement the JSON output function\n\nVERIFY: `cargo test timeline_json -- --nocapture`\n\n## Edge Cases\n- ISO 8601 conversion from ms epoch — ensure UTC timezone marker (\"Z\") is included\n- Empty events array → event_count: 0, events: [] (not omitted)\n- No expanded entities → expanded_entities: [] (not omitted)\n- serde_json::to_string should never fail on these types — unwrap is safe","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:28.374690Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:47:16.661533Z","compaction_level":0,"original_size":0,"labels":["cli","gate-3","phase-b","robot-mode"],"dependencies":[{"issue_id":"bd-dty","depends_on_id":"bd-3as","type":"blocks","created_at":"2026-02-02T21:33:37.703617Z","created_by":"tayloreernisse"},{"issue_id":"bd-dty","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:28.377349Z","created_by":"tayloreernisse"}]}
 {"id":"bd-epj","title":"[CP0] Config loading with Zod validation","description":"## Background\n\nConfig loading is critical infrastructure - every CLI command needs the config. Uses Zod for schema validation with sensible defaults. Must handle missing files gracefully with typed errors.\n\nReference: docs/prd/checkpoint-0.md sections \"Configuration Schema\", \"Config Resolution Order\"\n\n## Approach\n\n**src/core/config.ts:**\n```typescript\nimport { z } from 'zod';\nimport { readFileSync } from 'node:fs';\nimport { ConfigNotFoundError, ConfigValidationError } from './errors';\nimport { getConfigPath } from './paths';\n\nexport const ConfigSchema = z.object({\n  gitlab: z.object({\n    baseUrl: z.string().url(),\n    tokenEnvVar: z.string().default('GITLAB_TOKEN'),\n  }),\n  projects: z.array(z.object({\n    path: z.string().min(1),\n  })).min(1),\n  sync: z.object({\n    backfillDays: z.number().int().positive().default(14),\n    staleLockMinutes: z.number().int().positive().default(10),\n    heartbeatIntervalSeconds: z.number().int().positive().default(30),\n    cursorRewindSeconds: z.number().int().nonnegative().default(2),\n    primaryConcurrency: z.number().int().positive().default(4),\n    dependentConcurrency: z.number().int().positive().default(2),\n  }).default({}),\n  storage: z.object({\n    dbPath: z.string().optional(),\n    backupDir: z.string().optional(),\n    compressRawPayloads: z.boolean().default(true),\n  }).default({}),\n  embedding: z.object({\n    provider: z.literal('ollama').default('ollama'),\n    model: z.string().default('nomic-embed-text'),\n    baseUrl: z.string().url().default('http://localhost:11434'),\n    concurrency: z.number().int().positive().default(4),\n  }).default({}),\n});\n\nexport type Config = z.infer<typeof ConfigSchema>;\n\nexport function loadConfig(cliOverride?: string): Config {\n  const path = getConfigPath(cliOverride);\n  // throws ConfigNotFoundError if missing\n  // throws ConfigValidationError if invalid\n}\n```\n\n## Acceptance Criteria\n\n- [ ] `loadConfig()` returns validated Config object\n- [ ] `loadConfig()` throws ConfigNotFoundError if file missing\n- [ ] `loadConfig()` throws ConfigValidationError with Zod errors if invalid\n- [ ] Empty optional fields get default values\n- [ ] projects array must have at least 1 item\n- [ ] gitlab.baseUrl must be valid URL\n- [ ] All number fields must be positive integers\n- [ ] tests/unit/config.test.ts passes (8 tests)\n\n## Files\n\nCREATE:\n- src/core/config.ts\n- tests/unit/config.test.ts\n- tests/fixtures/mock-responses/valid-config.json\n- tests/fixtures/mock-responses/invalid-config.json\n\n## TDD Loop\n\nRED:\n```typescript\n// tests/unit/config.test.ts\ndescribe('Config', () => {\n  it('loads config from file path')\n  it('throws ConfigNotFoundError if file missing')\n  it('throws ConfigValidationError if required fields missing')\n  it('validates project paths are non-empty strings')\n  it('applies default values for optional fields')\n  it('loads from XDG path by default')\n  it('respects GI_CONFIG_PATH override')\n  it('respects --config flag override')\n})\n```\n\nGREEN: Implement loadConfig() function\n\nVERIFY: `npm run test -- tests/unit/config.test.ts`\n\n## Edge Cases\n\n- JSON parse error should wrap in ConfigValidationError\n- Zod error messages should be human-readable\n- File exists but empty → ConfigValidationError\n- File has extra fields → should pass (Zod strips by default)","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:49.091078Z","created_by":"tayloreernisse","updated_at":"2026-01-25T03:04:32.592139Z","closed_at":"2026-01-25T03:04:32.592003Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-epj","depends_on_id":"bd-gg1","type":"blocks","created_at":"2026-01-24T16:13:07.835800Z","created_by":"tayloreernisse"}]}
 {"id":"bd-gg1","title":"[CP0] Core utilities - paths, time, errors, logger","description":"## Background\n\nCore utilities provide the foundation for all other modules. Path resolution enables XDG-compliant config/data locations. Time utilities ensure consistent timestamp handling (ms epoch for DB, ISO for API). Error classes provide typed exceptions for clean error handling. Logger provides structured logging to stderr.\n\nReference: docs/prd/checkpoint-0.md sections \"Config + Data Locations\", \"Timestamp Convention\", \"Error Classes\", \"Logging Configuration\"\n\n## Approach\n\n**src/core/paths.ts:**\n- `getConfigPath(cliOverride?)`: resolution order is CLI flag → GI_CONFIG_PATH env → XDG default → local fallback\n- `getDataDir()`: uses XDG_DATA_HOME or ~/.local/share/gi\n- `getDbPath(configOverride?)`: returns data dir + data.db\n- `getBackupDir(configOverride?)`: returns data dir + backups/\n\n**src/core/time.ts:**\n- `isoToMs(isoString)`: converts GitLab API ISO 8601 → ms epoch\n- `msToIso(ms)`: converts ms epoch → ISO 8601\n- `nowMs()`: returns Date.now() for DB storage\n\n**src/core/errors.ts:**\nError hierarchy (all extend GiError base class with code and cause):\n- ConfigNotFoundError, ConfigValidationError\n- GitLabAuthError, GitLabNotFoundError, GitLabRateLimitError, GitLabNetworkError\n- DatabaseLockError, MigrationError\n- TokenNotSetError\n\n**src/core/logger.ts:**\n- pino logger to stderr (fd 2) with pino-pretty in dev\n- Child loggers: dbLogger, gitlabLogger, configLogger\n- LOG_LEVEL env var support (default: info)\n\n## Acceptance Criteria\n\n- [ ] `getConfigPath()` returns ~/.config/gi/config.json when no overrides\n- [ ] `getConfigPath()` respects GI_CONFIG_PATH env var\n- [ ] `getConfigPath(\"./custom.json\")` returns \"./custom.json\"\n- [ ] `isoToMs(\"2024-01-27T00:00:00.000Z\")` returns 1706313600000\n- [ ] `msToIso(1706313600000)` returns \"2024-01-27T00:00:00.000Z\"\n- [ ] All error classes have correct code property\n- [ ] Logger outputs to stderr (not stdout)\n- [ ] tests/unit/paths.test.ts passes\n- [ ] tests/unit/errors.test.ts passes\n\n## Files\n\nCREATE:\n- src/core/paths.ts\n- src/core/time.ts\n- src/core/errors.ts\n- src/core/logger.ts\n- tests/unit/paths.test.ts\n- tests/unit/errors.test.ts\n\n## TDD Loop\n\nRED: Write tests first\n```typescript\n// tests/unit/paths.test.ts\ndescribe('getConfigPath', () => {\n  it('uses XDG_CONFIG_HOME if set')\n  it('falls back to ~/.config/gi if XDG not set')\n  it('prefers --config flag over environment')\n  it('prefers environment over XDG default')\n  it('falls back to local gi.config.json in dev')\n})\n```\n\nGREEN: Implement paths.ts, errors.ts, time.ts, logger.ts\n\nVERIFY: `npm run test -- tests/unit/paths.test.ts tests/unit/errors.test.ts`\n\n## Edge Cases\n\n- XDG_CONFIG_HOME may not exist - don't create, just return path\n- existsSync() check for local fallback - only return if file exists\n- Time conversion must handle timezone edge cases - always use UTC\n- Logger must work even if pino-pretty not installed (production)","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-24T16:09:48.604382Z","created_by":"tayloreernisse","updated_at":"2026-01-25T02:53:26.527997Z","closed_at":"2026-01-25T02:53:26.527862Z","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-gg1","depends_on_id":"bd-327","type":"blocks","created_at":"2026-01-24T16:13:07.368187Z","created_by":"tayloreernisse"}]}
 {"id":"bd-hbo","title":"[CP1] Discussion ingestion module","description":"## Background\n\nDiscussion ingestion fetches all discussions and notes for a single issue. It is called as part of dependent sync - only for issues whose `updated_at` has advanced beyond `discussions_synced_for_updated_at`. After successful sync, it updates the watermark to prevent redundant refetches.\n\n## Approach\n\n### Module: src/ingestion/discussions.rs\n\n### Key Structs\n\n```rust\n#[derive(Debug, Default)]\npub struct IngestDiscussionsResult {\n    pub discussions_fetched: usize,\n    pub discussions_upserted: usize,\n    pub notes_upserted: usize,\n    pub system_notes_count: usize,\n}\n```\n\n### Main Function\n\n```rust\npub async fn ingest_issue_discussions(\n    conn: &Connection,\n    client: &GitLabClient,\n    config: &Config,\n    project_id: i64,           // Local DB project ID\n    gitlab_project_id: i64,    // GitLab project ID\n    issue_iid: i64,\n    local_issue_id: i64,\n    issue_updated_at: i64,     // For watermark update\n) -> Result<IngestDiscussionsResult>\n```\n\n### Logic\n\n1. Stream discussions via `client.paginate_issue_discussions()`\n2. For each discussion:\n   - Begin transaction\n   - Store raw payload (compressed based on config)\n   - Transform to NormalizedDiscussion\n   - Upsert discussion\n   - Get local discussion ID\n   - Transform notes via `transform_notes()`\n   - For each note: store raw payload, upsert note\n   - Track system_notes_count\n   - Commit transaction\n3. After all discussions processed: `mark_discussions_synced(conn, local_issue_id, issue_updated_at)`\n\n### Helper Functions\n\n```rust\nfn upsert_discussion(conn, discussion, payload_id) -> Result<()>\nfn get_local_discussion_id(conn, project_id, gitlab_id) -> Result<i64>\nfn upsert_note(conn, discussion_id, note, payload_id) -> Result<()>\nfn mark_discussions_synced(conn, issue_id, issue_updated_at) -> Result<()>\n```\n\n### Critical Invariant\n\n`discussions_synced_for_updated_at` MUST be updated only AFTER all discussions are successfully synced. This watermark prevents redundant refetches on subsequent runs.\n\n## Acceptance Criteria\n\n- [ ] `ingest_issue_discussions` streams all discussions for an issue\n- [ ] Each discussion wrapped in transaction for atomicity\n- [ ] Raw payloads stored for discussions and notes\n- [ ] `discussions_synced_for_updated_at` updated after successful sync\n- [ ] System notes tracked in result.system_notes_count\n- [ ] Notes linked to correct discussion via local discussion ID\n\n## Files\n\n- src/ingestion/mod.rs (add `pub mod discussions;`)\n- src/ingestion/discussions.rs (create)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/discussion_watermark_tests.rs\n#[tokio::test] async fn fetches_discussions_when_updated_at_advanced()\n#[tokio::test] async fn updates_watermark_after_successful_discussion_sync()\n#[tokio::test] async fn does_not_update_watermark_on_discussion_sync_failure()\n#[tokio::test] async fn stores_raw_payload_for_each_discussion()\n#[tokio::test] async fn stores_raw_payload_for_each_note()\n```\n\nGREEN: Implement ingest_issue_discussions with watermark logic\n\nVERIFY: `cargo test discussion_watermark`\n\n## Edge Cases\n\n- Issue with 0 discussions - mark synced anyway (empty is valid)\n- Discussion with 0 notes - should not happen per GitLab API (discussions always have >= 1 note)\n- Network failure mid-sync - watermark NOT updated, next run retries\n- individual_note=true discussions - have exactly 1 note","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.267582Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:52:47.500700Z","closed_at":"2026-01-25T22:52:47.500644Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-hbo","depends_on_id":"bd-1qf","type":"blocks","created_at":"2026-01-25T17:04:05.534265Z","created_by":"tayloreernisse"},{"issue_id":"bd-hbo","depends_on_id":"bd-2iq","type":"blocks","created_at":"2026-01-25T17:04:05.499474Z","created_by":"tayloreernisse"},{"issue_id":"bd-hbo","depends_on_id":"bd-xhz","type":"blocks","created_at":"2026-01-25T17:04:05.559260Z","created_by":"tayloreernisse"}]}
 {"id":"bd-hrs","title":"Create migration 007_documents.sql","description":"## Background\nMigration 007 creates the document storage layer that Gate A's entire search pipeline depends on. It introduces 5 tables: `documents` (the searchable unit), `document_labels` and `document_paths` (for filtered search), and two queue tables (`dirty_sources`, `pending_discussion_fetches`) that drive incremental document regeneration and discussion fetching in Gate C. This is the most-depended-on bead in the project (6 downstream beads block on it).\n\n## Approach\nCreate `migrations/007_documents.sql` with the exact SQL from PRD Section 1.1. The schema is fully specified in the PRD — no design decisions remain.\n\nKey implementation details:\n- `documents` table has `UNIQUE(source_type, source_id)` constraint for upsert support\n- `document_labels` and `document_paths` use `WITHOUT ROWID` for compact storage\n- `dirty_sources` uses composite PK `(source_type, source_id)` with `ON CONFLICT` upsert semantics\n- `pending_discussion_fetches` uses composite PK `(project_id, noteable_type, noteable_iid)`\n- Both queue tables have `next_attempt_at` indexed for efficient backoff queries\n- `labels_hash` and `paths_hash` on documents enable write optimization (skip unchanged labels/paths)\n\nRegister the migration in `src/core/db.rs` by adding entry 7 to the `MIGRATIONS` array.\n\n## Acceptance Criteria\n- [ ] `migrations/007_documents.sql` file exists with all 5 CREATE TABLE statements\n- [ ] Migration applies cleanly on fresh DB (`cargo test migration_tests`)\n- [ ] Migration applies cleanly after CP2 schema (migrations 001-006 already applied)\n- [ ] All foreign keys enforced: `documents.project_id -> projects(id)`, `document_labels.document_id -> documents(id) ON DELETE CASCADE`, `document_paths.document_id -> documents(id) ON DELETE CASCADE`, `pending_discussion_fetches.project_id -> projects(id)`\n- [ ] All indexes created: `idx_documents_project_updated`, `idx_documents_author`, `idx_documents_source`, `idx_documents_hash`, `idx_document_labels_label`, `idx_document_paths_path`, `idx_dirty_sources_next_attempt`, `idx_pending_discussions_next_attempt`\n- [ ] `labels_hash TEXT NOT NULL DEFAULT ''` and `paths_hash TEXT NOT NULL DEFAULT ''` columns present on `documents`\n- [ ] Schema version 7 recorded in `schema_version` table\n- [ ] `cargo build` succeeds after registering migration in db.rs\n\n## Files\n- `migrations/007_documents.sql` — new file (copy exact SQL from PRD Section 1.1)\n- `src/core/db.rs` — add migration 7 to `MIGRATIONS` array\n\n## TDD Loop\nRED: Add migration to db.rs, run `cargo test migration_tests` — fails because SQL file missing\nGREEN: Create `migrations/007_documents.sql` with full schema\nVERIFY: `cargo test migration_tests && cargo build`\n\n## Edge Cases\n- Migration must be idempotent-safe if applied twice (INSERT into schema_version will fail on second run — this is expected and handled by the migration runner's version check)\n- `WITHOUT ROWID` tables (document_labels, document_paths) require explicit PK — already defined\n- `CHECK` constraint on `documents.source_type` must match exactly: `'issue','merge_request','discussion'`\n- `CHECK` constraint on `documents.truncated_reason` allows NULL or one of 4 specific values","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:25:25.734380Z","created_by":"tayloreernisse","updated_at":"2026-01-30T16:54:12.854351Z","closed_at":"2026-01-30T16:54:12.854149Z","close_reason":"Completed: migration 007_documents.sql with 5 tables (documents, document_labels, document_paths, dirty_sources, pending_discussion_fetches), 8 indexes, registered in db.rs, cargo build + migration tests pass","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-hrs","depends_on_id":"bd-3lc","type":"blocks","created_at":"2026-01-30T15:29:15.536304Z","created_by":"tayloreernisse"}]}
+{"id":"bd-hu3","title":"Write migration 011: resource event tables, entity_references, and dependent fetch queue","description":"## Background\nPhase B needs three new event tables and a generic dependent fetch queue to power temporal queries (timeline, file-history, trace). These tables store structured event data from GitLab Resource Events APIs, replacing fragile system note parsing for state/label/milestone changes.\n\nMigration 010_chunk_config.sql already exists, so Phase B starts at migration 011.\n\n## Approach\nCreate migrations/011_resource_events.sql with the exact schema from the Phase B spec (§1.2 + §2.2):\n\n**Event tables:**\n- resource_state_events: state changes (opened/closed/reopened/merged/locked) with source_merge_request_id for \"closed by MR\" linking\n- resource_label_events: label add/remove with label_name\n- resource_milestone_events: milestone add/remove with milestone_title + milestone_id\n\n**Cross-reference table (Gate 2):**\n- entity_references: source/target entity pairs with reference_type (closes/mentioned/related), source_method provenance, and unresolved reference support (target_entity_id NULL with target_project_path + target_entity_iid)\n\n**Dependent fetch queue:**\n- pending_dependent_fetches: generic job queue with job_type IN ('resource_events', 'mr_closes_issues', 'mr_diffs'), locked_at crash recovery, exponential backoff via attempts + next_retry_at\n\n**All tables must have:**\n- CHECK constraints for entity exclusivity (issue XOR merge_request) on event tables\n- UNIQUE constraints (gitlab_id + project_id for events, composite for queue, multi-column for references)\n- Partial indexes (WHERE issue_id IS NOT NULL, WHERE target_entity_id IS NULL, etc.)\n- CASCADE deletes on project_id and entity FKs\n\nRegister in src/core/db.rs MIGRATIONS array:\n```rust\n(\"011\", include_str!(\"../../migrations/011_resource_events.sql\")),\n```\n\nEnd migration with:\n```sql\nINSERT INTO schema_version (version, applied_at, description)\nVALUES (11, strftime('%s', 'now') * 1000, 'Resource events, entity references, and dependent fetch queue');\n```\n\n## Acceptance Criteria\n- [ ] migrations/011_resource_events.sql exists with all 4 tables + indexes + constraints\n- [ ] src/core/db.rs MIGRATIONS array includes (\"011\", include_str!(...))\n- [ ] `cargo build` succeeds (migration SQL compiles into binary)\n- [ ] `cargo test migration` passes (migration applies cleanly on fresh DB)\n- [ ] All CHECK constraints enforced (issue XOR merge_request on event tables)\n- [ ] All UNIQUE constraints present (prevents duplicate events/refs/jobs)\n- [ ] entity_references UNIQUE handles NULL coalescing correctly\n- [ ] pending_dependent_fetches job_type CHECK includes all three types\n\n## Files\n- migrations/011_resource_events.sql (new)\n- src/core/db.rs (add to MIGRATIONS array, line ~46)\n\n## TDD Loop\nRED: Add test in tests/migration_tests.rs:\n- `test_migration_011_creates_event_tables` - verify all 4 tables exist after migration\n- `test_migration_011_entity_exclusivity_constraint` - verify CHECK rejects both NULL and both non-NULL for issue_id/merge_request_id\n- `test_migration_011_event_dedup` - verify UNIQUE(gitlab_id, project_id) rejects duplicate events\n- `test_migration_011_entity_references_dedup` - verify UNIQUE constraint with NULL coalescing\n- `test_migration_011_queue_dedup` - verify UNIQUE(project_id, entity_type, entity_iid, job_type)\n\nGREEN: Write the migration SQL + register in db.rs\n\nVERIFY: `cargo test migration_tests -- --nocapture`\n\n## Edge Cases\n- entity_references UNIQUE uses COALESCE for NULLable columns — test with both resolved and unresolved refs\n- pending_dependent_fetches job_type CHECK — ensure 'mr_diffs' is included (Gate 4 needs it)\n- SQLite doesn't enforce CHECK on INSERT OR REPLACE — verify constraint behavior\n- The entity exclusivity CHECK must allow exactly one of issue_id/merge_request_id to be non-NULL\n- Verify CASCADE deletes work (delete project → all events/refs/jobs deleted)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:23.933894Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:40:56.615120Z","compaction_level":0,"original_size":0,"labels":["gate-1","phase-b","schema"],"dependencies":[{"issue_id":"bd-hu3","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:23.937375Z","created_by":"tayloreernisse"}]}
 {"id":"bd-iba","title":"Add GitLab client MR pagination methods","description":"## Background\nGitLab client pagination for merge requests and discussions. Must support robust pagination with fallback chain because some GitLab instances/proxies strip headers.\n\n## Approach\nAdd to existing `src/gitlab/client.rs`:\n1. `MergeRequestPage` struct - Items + pagination metadata\n2. `parse_link_header_next()` - RFC 8288 Link header parsing\n3. `fetch_merge_requests_page()` - Single page fetch with metadata\n4. `paginate_merge_requests()` - Async stream for all MRs\n5. `paginate_mr_discussions()` - Async stream for MR discussions\n\n## Files\n- `src/gitlab/client.rs` - Add pagination methods\n\n## Acceptance Criteria\n- [ ] `MergeRequestPage` struct exists with `items`, `next_page`, `is_last_page`\n- [ ] `parse_link_header_next()` extracts `rel=\"next\"` URL from Link header\n- [ ] Pagination fallback chain: Link header > x-next-page > full-page heuristic\n- [ ] `paginate_merge_requests()` returns `Pin<Box<dyn Stream<Item = Result<GitLabMergeRequest>>>>`\n- [ ] `paginate_mr_discussions()` returns `Pin<Box<dyn Stream<Item = Result<GitLabDiscussion>>>>`\n- [ ] MR endpoint uses `scope=all&state=all` to include all MRs\n- [ ] `cargo test client` passes\n\n## TDD Loop\nRED: `cargo test fetch_merge_requests` -> method not found\nGREEN: Add pagination methods\nVERIFY: `cargo test client`\n\n## Struct Definitions\n```rust\n#[derive(Debug)]\npub struct MergeRequestPage {\n    pub items: Vec<GitLabMergeRequest>,\n    pub next_page: Option<u32>,\n    pub is_last_page: bool,\n}\n```\n\n## Link Header Parsing (RFC 8288)\n```rust\n/// Parse Link header to extract rel=\"next\" URL.\nfn parse_link_header_next(headers: &reqwest::header::HeaderMap) -> Option<String> {\n    headers\n        .get(\"link\")\n        .and_then(|v| v.to_str().ok())\n        .and_then(|link_str| {\n            // Format: <url>; rel=\"next\", <url>; rel=\"last\"\n            for part in link_str.split(',') {\n                let part = part.trim();\n                if part.contains(\"rel=\\\"next\\\"\") || part.contains(\"rel=next\") {\n                    if let Some(start) = part.find('<') {\n                        if let Some(end) = part.find('>') {\n                            return Some(part[start + 1..end].to_string());\n                        }\n                    }\n                }\n            }\n            None\n        })\n}\n```\n\n## Pagination Fallback Chain\n```rust\nlet next_page = match (link_next, x_next_page, items.len() as u32 == per_page) {\n    (Some(_), _, _) => Some(page + 1),     // Link header present: continue\n    (None, Some(np), _) => Some(np),       // x-next-page present: use it\n    (None, None, true) => Some(page + 1),  // Full page, no headers: try next\n    (None, None, false) => None,           // Partial page: we're done\n};\n```\n\n## Fetch Single Page\n```rust\npub async fn fetch_merge_requests_page(\n    &self,\n    gitlab_project_id: i64,\n    updated_after: Option<i64>,\n    cursor_rewind_seconds: u32,\n    page: u32,\n    per_page: u32,\n) -> Result<MergeRequestPage> {\n    let mut params = vec![\n        (\"scope\", \"all\".to_string()),\n        (\"state\", \"all\".to_string()),\n        (\"order_by\", \"updated_at\".to_string()),\n        (\"sort\", \"asc\".to_string()),\n        (\"per_page\", per_page.to_string()),\n        (\"page\", page.to_string()),\n    ];\n    // Apply cursor rewind for safety\n    // ...\n}\n```\n\n## Async Stream Pattern\n```rust\npub fn paginate_merge_requests(\n    &self,\n    gitlab_project_id: i64,\n    updated_after: Option<i64>,\n    cursor_rewind_seconds: u32,\n) -> Pin<Box<dyn Stream<Item = Result<GitLabMergeRequest>> + Send + '_>> {\n    Box::pin(async_stream::try_stream! {\n        let mut page = 1u32;\n        let per_page = 100u32;\n        loop {\n            let page_result = self.fetch_merge_requests_page(...).await?;\n            for mr in page_result.items {\n                yield mr;\n            }\n            if page_result.is_last_page {\n                break;\n            }\n            match page_result.next_page {\n                Some(np) => page = np,\n                None => break,\n            }\n        }\n    })\n}\n```\n\n## Edge Cases\n- `scope=all` required to include all MRs (not just authored by current user)\n- `state=all` required to include merged/closed (GitLab defaults may exclude)\n- `locked` state cannot be filtered server-side (use local SQL filtering)\n- Cursor rewind should clamp to 0 to avoid negative timestamps","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:41.633065Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:13:05.613625Z","closed_at":"2026-01-27T00:13:05.613440Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-iba","depends_on_id":"bd-5ta","type":"blocks","created_at":"2026-01-26T22:08:54.364647Z","created_by":"tayloreernisse"}]}
+{"id":"bd-ike","title":"Epic: Gate 3 - Decision Timeline (lore timeline)","description":"Implement 'lore timeline <query>' command that produces keyword-driven chronological narratives across issues, MRs, and discussions. Combines FTS5 search (seed), BFS cross-reference expansion, event collection from resource event tables, and evidence note surfacing. Output in both human-readable and robot JSON formats.\n\nChildren: bd-20e (types), bd-32q (seed phase), bd-ypa (expand phase), bd-3as (collect events), bd-2f2 (human output), bd-dty (robot JSON), bd-1nf (CLI wiring)","status":"open","priority":1,"issue_type":"feature","created_at":"2026-02-02T21:31:01.036474Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:50:48.982831Z","compaction_level":0,"original_size":0,"labels":["epic","gate-3","phase-b"],"dependencies":[{"issue_id":"bd-ike","depends_on_id":"bd-1se","type":"blocks","created_at":"2026-02-02T21:33:37.875622Z","created_by":"tayloreernisse"},{"issue_id":"bd-ike","depends_on_id":"bd-2zl","type":"blocks","created_at":"2026-02-02T21:33:37.831914Z","created_by":"tayloreernisse"}]}
+{"id":"bd-jec","title":"Add fetchMrFileChanges config flag","description":"## Background\nMR file change fetching should be opt-in (default true) to avoid extra API calls. Follows the same pattern as fetchResourceEvents (bd-2e8).\n\n## Approach\nAdd to SyncConfig in src/core/config.rs:\n```rust\n#[serde(rename = \"fetchMrFileChanges\", default = \"default_true\")]\npub fetch_mr_file_changes: bool,\n```\n\nUpdate Default impl for SyncConfig to include `fetch_mr_file_changes: true`.\n\nAdd --no-file-changes CLI flag to sync command in src/cli/mod.rs (SyncArgs):\n```rust\n/// Skip MR file change fetching (overrides config)\n#[arg(long = \"no-file-changes\", help_heading = \"Sync Options\")]\npub no_file_changes: bool,\n```\n\nIn sync handler, override config:\n```rust\nif args.no_file_changes {\n    config.sync.fetch_mr_file_changes = false;\n}\n```\n\n## Acceptance Criteria\n- [ ] SyncConfig deserializes fetchMrFileChanges from JSON config\n- [ ] Defaults to true when field absent\n- [ ] --no-file-changes flag parses correctly\n- [ ] --no-file-changes overrides config to false\n- [ ] `cargo test` passes\n\n## Files\n- src/core/config.rs (add field + Default update)\n- src/cli/mod.rs (add --no-file-changes to SyncArgs)\n- src/cli/commands/sync.rs (override logic)\n\n## TDD Loop\nRED: tests/config_tests.rs:\n- `test_sync_config_fetch_mr_file_changes_default_true`\n- `test_sync_config_fetch_mr_file_changes_explicit_false`\n\nGREEN: Add field, default, flag, override\n\nVERIFY: `cargo test config -- --nocapture && cargo build`\n\n## Edge Cases\n- Same as bd-2e8: reuse default_true fn, check for naming conflicts with existing flags","status":"open","priority":3,"issue_type":"task","created_at":"2026-02-02T21:34:08.892666Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:48:11.012962Z","compaction_level":0,"original_size":0,"labels":["config","gate-4","phase-b"],"dependencies":[{"issue_id":"bd-jec","depends_on_id":"bd-14q","type":"parent-child","created_at":"2026-02-02T21:34:08.895167Z","created_by":"tayloreernisse"}]}
 {"id":"bd-jov","title":"[CP1] Discussion and note transformers","description":"Transform GitLab discussion/note payloads to normalized database schema.\n\n## Module\nsrc/gitlab/transformers/discussion.rs\n\n## Structs\n\n### NormalizedDiscussion\n- gitlab_discussion_id: String\n- project_id: i64\n- issue_id: i64\n- noteable_type: String (\"Issue\")\n- individual_note: bool\n- first_note_at, last_note_at: Option<i64>\n- last_seen_at: i64\n- resolvable, resolved: bool\n\n### NormalizedNote\n- gitlab_id: i64\n- project_id: i64\n- note_type: Option<String>\n- is_system: bool\n- author_username: String\n- body: String\n- created_at, updated_at, last_seen_at: i64\n- position: i32 (array index in notes[])\n- resolvable, resolved: bool\n- resolved_by: Option<String>\n- resolved_at: Option<i64>\n\n## Functions\n\n### transform_discussion(gitlab_discussion, local_project_id, local_issue_id) -> NormalizedDiscussion\n- Compute first_note_at/last_note_at from notes array min/max created_at\n- Compute resolvable (any note resolvable)\n- Compute resolved (resolvable AND all resolvable notes resolved)\n\n### transform_notes(gitlab_discussion, local_project_id) -> Vec<NormalizedNote>\n- Enumerate notes to get position (array index)\n- Set is_system from note.system\n- Convert timestamps to ms epoch\n\nFiles: src/gitlab/transformers/discussion.rs\nTests: tests/discussion_transformer_tests.rs\nDone when: Unit tests pass for discussion/note transformation with system note flagging","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T15:43:04.481361Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.759691Z","deleted_at":"2026-01-25T17:02:01.759684Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-k7b","title":"[CP1] gi show issue command","description":"Show issue details with discussions.\n\n## Module\nsrc/cli/commands/show.rs\n\n## Clap Definition\nShow {\n    #[arg(value_parser = [\"issue\", \"mr\"])]\n    entity: String,\n    \n    iid: i64,\n    \n    #[arg(long)]\n    project: Option<String>,\n}\n\n## Output Format\nIssue #1234: Authentication redesign\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\nProject:  group/project-one\nState:    opened\nAuthor:   @johndoe\nCreated:  2024-01-15\nUpdated:  2024-03-20\nLabels:   enhancement, auth\nURL:      https://gitlab.example.com/group/project-one/-/issues/1234\n\nDescription:\n  We need to redesign the authentication flow to support...\n\nDiscussions (5):\n\n  @janedoe (2024-01-16):\n    I agree we should move to JWT-based auth...\n\n    @johndoe (2024-01-16):\n      What about refresh token strategy?\n\n  @bobsmith (2024-01-17):\n    Have we considered OAuth2?\n\n## Ambiguity Handling\nIf multiple projects have same iid, either:\n- Prompt for --project flag\n- Show error listing which projects have that iid\n\nFiles: src/cli/commands/show.rs\nDone when: Issue detail view displays all fields including threaded discussions","status":"tombstone","priority":3,"issue_type":"task","created_at":"2026-01-25T16:58:26.904813Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:01.944183Z","deleted_at":"2026-01-25T17:02:01.944179Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
 {"id":"bd-lcb","title":"Epic: CP2 Gate E - CLI Complete","description":"## Background\nGate E validates all CLI commands are functional and user-friendly. This is the final usability gate - even if all data is correct, users need good CLI UX to access it.\n\n## Acceptance Criteria (Pass/Fail)\n\n### List Command\n- [ ] `gi list mrs` shows MR table with columns: iid, title, state, author, branches, updated\n- [ ] `gi list mrs --state=opened` filters to only opened MRs\n- [ ] `gi list mrs --state=merged` filters to only merged MRs\n- [ ] `gi list mrs --state=closed` filters to only closed MRs\n- [ ] `gi list mrs --state=locked` filters locally (not server-side filter)\n- [ ] `gi list mrs --draft` shows only draft MRs\n- [ ] `gi list mrs --no-draft` excludes draft MRs\n- [ ] Draft MRs show `[DRAFT]` prefix in title column\n- [ ] `gi list mrs --author=username` filters by author\n- [ ] `gi list mrs --assignee=username` filters by assignee\n- [ ] `gi list mrs --reviewer=username` filters by reviewer\n- [ ] `gi list mrs --target-branch=main` filters by target branch\n- [ ] `gi list mrs --source-branch=feature/x` filters by source branch\n- [ ] `gi list mrs --label=bugfix` filters by label\n- [ ] `gi list mrs --limit=N` limits output\n\n### Show Command\n- [ ] `gi show mr <iid>` displays full MR detail\n- [ ] Show includes: title, description, state, draft status, author\n- [ ] Show includes: assignees, reviewers, labels\n- [ ] Show includes: source_branch, target_branch\n- [ ] Show includes: detailed_merge_status (e.g., \"mergeable\")\n- [ ] Show includes: merge_user and merged_at for merged MRs\n- [ ] Show includes: discussions with author and date\n- [ ] DiffNote shows file context: `[src/file.ts:45]`\n- [ ] Multi-line DiffNote shows range: `[src/file.ts:45-48]`\n- [ ] Resolved discussions show `[RESOLVED]` marker\n\n### Count Command\n- [ ] `gi count mrs` shows total count\n- [ ] Count shows state breakdown: opened, merged, closed\n\n### Sync Status\n- [ ] `gi sync-status` shows MR cursor position\n- [ ] Sync status shows last sync timestamp\n\n## Validation Script\n```bash\n#!/bin/bash\nset -e\n\nDB_PATH=\"${XDG_DATA_HOME:-$HOME/.local/share}/gitlab-inbox/db.sqlite3\"\n\necho \"=== Gate E: CLI Complete ===\"\n\n# 1. Test list command (basic)\necho \"Step 1: Basic list...\"\ngi list mrs --limit=5 || { echo \"FAIL: list mrs failed\"; exit 1; }\n\n# 2. Test state filters\necho \"Step 2: State filters...\"\nfor state in opened merged closed; do\n  echo \"  Testing --state=$state\"\n  gi list mrs --state=$state --limit=3 || echo \"  Warning: No $state MRs\"\ndone\n\n# 3. Test draft filters\necho \"Step 3: Draft filters...\"\ngi list mrs --draft --limit=3 || echo \"  Note: No draft MRs found\"\ngi list mrs --no-draft --limit=3 || echo \"  Note: All MRs are drafts?\"\n\n# 4. Check [DRAFT] prefix\necho \"Step 4: Check [DRAFT] prefix...\"\nDRAFT_IID=$(sqlite3 \"$DB_PATH\" \"SELECT iid FROM merge_requests WHERE draft = 1 LIMIT 1;\")\nif [ -n \"$DRAFT_IID\" ]; then\n  if gi list mrs --limit=100 | grep -q \"\\[DRAFT\\]\"; then\n    echo \"  PASS: [DRAFT] prefix found\"\n  else\n    echo \"  FAIL: Draft MR exists but no [DRAFT] prefix in output\"\n  fi\nelse\n  echo \"  Skip: No draft MRs to test\"\nfi\n\n# 5. Test author/assignee/reviewer filters\necho \"Step 5: User filters...\"\nAUTHOR=$(sqlite3 \"$DB_PATH\" \"SELECT author_username FROM merge_requests LIMIT 1;\")\nif [ -n \"$AUTHOR\" ]; then\n  echo \"  Testing --author=$AUTHOR\"\n  gi list mrs --author=\"$AUTHOR\" --limit=3\nfi\n\nREVIEWER=$(sqlite3 \"$DB_PATH\" \"SELECT username FROM mr_reviewers LIMIT 1;\")\nif [ -n \"$REVIEWER\" ]; then\n  echo \"  Testing --reviewer=$REVIEWER\"\n  gi list mrs --reviewer=\"$REVIEWER\" --limit=3\nfi\n\n# 6. Test branch filters\necho \"Step 6: Branch filters...\"\nTARGET=$(sqlite3 \"$DB_PATH\" \"SELECT target_branch FROM merge_requests LIMIT 1;\")\nif [ -n \"$TARGET\" ]; then\n  echo \"  Testing --target-branch=$TARGET\"\n  gi list mrs --target-branch=\"$TARGET\" --limit=3\nfi\n\n# 7. Test show command\necho \"Step 7: Show command...\"\nMR_IID=$(sqlite3 \"$DB_PATH\" \"SELECT iid FROM merge_requests LIMIT 1;\")\ngi show mr \"$MR_IID\" || { echo \"FAIL: show mr failed\"; exit 1; }\n\n# 8. Test show with DiffNote context\necho \"Step 8: Show with DiffNote...\"\nDIFFNOTE_MR=$(sqlite3 \"$DB_PATH\" \"\n  SELECT DISTINCT m.iid\n  FROM merge_requests m\n  JOIN discussions d ON d.merge_request_id = m.id\n  JOIN notes n ON n.discussion_id = d.id\n  WHERE n.position_new_path IS NOT NULL\n  LIMIT 1;\n\")\nif [ -n \"$DIFFNOTE_MR\" ]; then\n  echo \"  Testing MR with DiffNotes: !$DIFFNOTE_MR\"\n  OUTPUT=$(gi show mr \"$DIFFNOTE_MR\")\n  if echo \"$OUTPUT\" | grep -qE '\\[[^]]+:[0-9]+\\]'; then\n    echo \"  PASS: File context [path:line] found\"\n  else\n    echo \"  FAIL: DiffNote should show [path:line] context\"\n  fi\nelse\n  echo \"  Skip: No MRs with DiffNotes\"\nfi\n\n# 9. Test count command\necho \"Step 9: Count command...\"\ngi count mrs || { echo \"FAIL: count mrs failed\"; exit 1; }\n\n# 10. Test sync-status\necho \"Step 10: Sync status...\"\ngi sync-status || echo \"  Note: sync-status may need implementation\"\n\necho \"\"\necho \"=== Gate E: PASSED ===\"\n```\n\n## Test Commands (Quick Verification)\n```bash\n# List with all column types visible:\ngi list mrs --limit=10\n\n# Show a specific MR:\ngi show mr 42\n\n# Count with breakdown:\ngi count mrs\n\n# Complex filter:\ngi list mrs --state=opened --reviewer=alice --target-branch=main --limit=5\n```\n\n## Expected Output Formats\n\n### gi list mrs\n```\nMerge Requests (showing 5 of 1,234)\n\n  !847  Refactor auth to use JWT tokens           merged   @johndoe   main <- feature/jwt   3d ago\n  !846  Fix memory leak in websocket handler      opened   @janedoe   main <- fix/websocket 5d ago\n  !845  [DRAFT] Add dark mode CSS variables       opened   @bobsmith  main <- ui/dark-mode  1w ago\n  !844  Update dependencies to latest versions    closed   @alice     main <- chore/deps    2w ago\n```\n\n### gi show mr 847\n```\nMerge Request !847: Refactor auth to use JWT tokens\n================================================================================\n\nProject:        group/project-one\nState:          merged\nDraft:          No\nAuthor:         @johndoe\nAssignees:      @janedoe, @bobsmith\nReviewers:      @alice, @charlie\nLabels:         enhancement, auth, reviewed\nSource:         feature/jwt\nTarget:         main\nMerge Status:   merged\nMerged By:      @alice\nMerged At:      2024-03-20 14:30:00\n\nDescription:\n  Moving away from session cookies to JWT-based authentication...\n\nDiscussions (3):\n\n  @janedoe (2024-03-16) [src/auth/jwt.ts:45]:\n    Should we use a separate signing key for refresh tokens?\n\n    @johndoe (2024-03-16):\n      Good point. I'll add a separate key with rotation support.\n\n  @alice (2024-03-18) [RESOLVED]:\n    Looks good! Just one nit about the token expiry constant.\n```\n\n### gi count mrs\n```\nMerge Requests: 1,234\n  opened: 89\n  merged: 1,045\n  closed: 100\n```\n\n## Dependencies\nThis gate requires:\n- bd-3js (CLI commands implementation)\n- All previous gates must pass first\n\n## Edge Cases\n- Ambiguous MR iid across projects: should prompt for `--project` or show error\n- Very long titles: should truncate with `...` in list view\n- Empty description: should show \"No description\" or empty section\n- No discussions: should show \"No discussions\" message\n- Unicode in titles/descriptions: should render correctly","status":"closed","priority":3,"issue_type":"task","created_at":"2026-01-26T22:06:02.411132Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:48:21.061166Z","closed_at":"2026-01-27T00:48:21.061125Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-lcb","depends_on_id":"bd-3js","type":"blocks","created_at":"2026-01-26T22:08:55.957747Z","created_by":"tayloreernisse"}]}
@@ -112,7 +144,11 @@
 {"id":"bd-ozy","title":"[CP1] Ingestion orchestrator","description":"## Background\n\nThe ingestion orchestrator coordinates issue sync followed by dependent discussion sync. It implements the CP1 canonical pattern: fetch issues, identify which need discussion sync (updated_at advanced), then execute discussion sync with bounded concurrency.\n\n## Approach\n\n### Module: src/ingestion/orchestrator.rs\n\n### Main Function\n\n```rust\npub async fn ingest_project_issues(\n    conn: &Connection,\n    client: &GitLabClient,\n    config: &Config,\n    project_id: i64,           // Local DB project ID\n    gitlab_project_id: i64,\n) -> Result<IngestProjectResult>\n\n#[derive(Debug, Default)]\npub struct IngestProjectResult {\n    pub issues_fetched: usize,\n    pub issues_upserted: usize,\n    pub labels_created: usize,\n    pub discussions_fetched: usize,\n    pub notes_fetched: usize,\n    pub system_notes_count: usize,\n    pub issues_skipped_discussion_sync: usize,\n}\n```\n\n### Orchestration Steps\n\n1. **Call issue ingestion**: `ingest_issues(conn, client, config, project_id, gitlab_project_id)`\n2. **Get issues needing discussion sync**: From IngestIssuesResult.issues_needing_discussion_sync\n3. **Execute bounded discussion sync**:\n   - Use `tokio::task::LocalSet` for single-threaded runtime\n   - Respect `config.sync.dependent_concurrency` (default: 5)\n   - For each IssueForDiscussionSync:\n     - Call `ingest_issue_discussions(...)`\n     - Aggregate results\n4. **Calculate skipped count**: total_issues - issues_needing_discussion_sync.len()\n\n### Bounded Concurrency Pattern\n\n```rust\nuse futures::stream::{self, StreamExt};\n\nlet local_set = LocalSet::new();\nlocal_set.run_until(async {\n    stream::iter(issues_needing_sync)\n        .map(|issue| async {\n            ingest_issue_discussions(\n                conn, client, config,\n                project_id, gitlab_project_id,\n                issue.iid, issue.local_issue_id, issue.updated_at,\n            ).await\n        })\n        .buffer_unordered(config.sync.dependent_concurrency)\n        .try_collect::<Vec<_>>()\n        .await\n}).await\n```\n\nNote: Single-threaded runtime means concurrency is I/O-bound, not parallel execution.\n\n## Acceptance Criteria\n\n- [ ] Orchestrator calls issue ingestion first\n- [ ] Only issues with updated_at > discussions_synced_for_updated_at get discussion sync\n- [ ] Bounded concurrency respects dependent_concurrency config\n- [ ] Results aggregated from both issue and discussion ingestion\n- [ ] issues_skipped_discussion_sync accurately reflects unchanged issues\n\n## Files\n\n- src/ingestion/mod.rs (add `pub mod orchestrator;`)\n- src/ingestion/orchestrator.rs (create)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/orchestrator_tests.rs\n#[tokio::test] async fn orchestrates_issue_then_discussion_sync()\n#[tokio::test] async fn skips_discussion_sync_for_unchanged_issues()\n#[tokio::test] async fn respects_bounded_concurrency()\n#[tokio::test] async fn aggregates_results_correctly()\n```\n\nGREEN: Implement orchestrator with bounded concurrency\n\nVERIFY: `cargo test orchestrator`\n\n## Edge Cases\n\n- All issues unchanged - no discussion sync calls\n- All issues new - all get discussion sync\n- dependent_concurrency=1 - sequential discussion fetches\n- Issue ingestion fails - orchestrator returns error, no discussion sync","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.289941Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:54:07.447647Z","closed_at":"2026-01-25T22:54:07.447577Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-ozy","depends_on_id":"bd-208","type":"blocks","created_at":"2026-01-25T17:04:05.583955Z","created_by":"tayloreernisse"},{"issue_id":"bd-ozy","depends_on_id":"bd-hbo","type":"blocks","created_at":"2026-01-25T17:04:05.605851Z","created_by":"tayloreernisse"}]}
 {"id":"bd-pr1","title":"Implement lore stats CLI command","description":"## Background\nThe stats command provides visibility into the document/search/embedding pipeline health. It reports counts (DocumentStats, EmbeddingStats, FtsStats, QueueStats), verifies consistency between tables (--check), and repairs inconsistencies (--repair). This is essential for diagnosing sync issues and validating Gate A/B/C correctness.\n\n## Approach\nCreate `src/cli/commands/stats.rs` per PRD Section 4.6.\n\n**Stats structs (PRD-exact):**\n```rust\n#[derive(Debug, Serialize)]\npub struct Stats {\n    pub documents: DocumentStats,\n    pub embeddings: EmbeddingStats,\n    pub fts: FtsStats,\n    pub queues: QueueStats,\n}\n\n#[derive(Debug, Serialize)]\npub struct DocumentStats {\n    pub issues: usize,\n    pub mrs: usize,\n    pub discussions: usize,\n    pub total: usize,\n    pub truncated: usize,\n}\n\n#[derive(Debug, Serialize)]\npub struct EmbeddingStats {\n    /// Documents with at least one embedding (chunk_index=0 exists in embedding_metadata)\n    pub embedded: usize,\n    pub pending: usize,\n    pub failed: usize,\n    /// embedded / total_documents * 100 (document-level, not chunk-level)\n    pub coverage_pct: f64,\n    /// Total chunks across all embedded documents\n    pub total_chunks: usize,\n}\n\n#[derive(Debug, Serialize)]\npub struct FtsStats { pub indexed: usize }\n\n#[derive(Debug, Serialize)]\npub struct QueueStats {\n    pub dirty_sources: usize,\n    pub dirty_sources_failed: usize,\n    pub pending_discussion_fetches: usize,\n    pub pending_discussion_fetches_failed: usize,\n}\n```\n\n**IntegrityCheck struct (PRD-exact):**\n```rust\n#[derive(Debug, Serialize)]\npub struct IntegrityCheck {\n    pub documents_count: usize,\n    pub fts_count: usize,\n    pub embeddings_count: usize,\n    pub metadata_count: usize,\n    pub orphaned_embeddings: usize,\n    pub hash_mismatches: usize,\n    pub ok: bool,\n}\n```\n\n**RepairResult struct (PRD-exact):**\n```rust\n#[derive(Debug, Serialize)]\npub struct RepairResult {\n    pub orphaned_embeddings_deleted: usize,\n    pub stale_embeddings_cleared: usize,\n    pub missing_fts_repopulated: usize,\n}\n```\n\n**Core functions:**\n- `run_stats(config) -> Result<Stats>` — gather all stats\n- `run_integrity_check(config) -> Result<IntegrityCheck>` — verify consistency\n- `run_repair(config) -> Result<RepairResult>` — fix issues\n\n**Integrity checks (per PRD):**\n1. documents count == documents_fts count\n2. All `embeddings.rowid / 1000` map to valid `documents.id` (orphan detection)\n3. `embedding_metadata.document_hash == documents.content_hash` for chunk_index=0 rows (staleness uses `document_hash`, NOT `chunk_hash`)\n\n**Repair operations (PRD-exact):**\n1. Delete orphaned embedding_metadata (document_id NOT IN documents)\n2. Delete orphaned vec0 rows: `DELETE FROM embeddings WHERE rowid / 1000 NOT IN (SELECT id FROM documents)` — uses `rowid / 1000` for chunked scheme\n3. Clear stale embeddings: find documents where `embedding_metadata.document_hash != documents.content_hash` (chunk_index=0 comparison), delete ALL chunks for those docs (range-based: `rowid >= doc_id * 1000 AND rowid < (doc_id + 1) * 1000`)\n4. FTS rebuild: `INSERT INTO documents_fts(documents_fts) VALUES('rebuild')` — full rebuild, NOT optimize. PRD note: partial fix is fragile with external-content FTS; rebuild is guaranteed correct.\n\n**CLI args (PRD-exact):**\n```rust\n#[derive(Args)]\npub struct StatsArgs {\n    #[arg(long)]\n    check: bool,\n    #[arg(long, requires = \"check\")]\n    repair: bool,  // --repair requires --check\n}\n```\n\n## Acceptance Criteria\n- [ ] Document counts by type: issues, mrs, discussions, total, truncated\n- [ ] Embedding coverage is document-level (not chunk-level): `embedded / total * 100`\n- [ ] Embedding stats include total_chunks count\n- [ ] FTS indexed count reported\n- [ ] Queue stats: dirty_sources + dirty_sources_failed, pending_discussion_fetches + pending_discussion_fetches_failed\n- [ ] --check verifies: FTS count == documents count, orphan embeddings, hash mismatches\n- [ ] Orphan detection uses `rowid / 1000` for chunked embedding scheme\n- [ ] Hash mismatch uses `document_hash` (not `chunk_hash`) for document-level staleness\n- [ ] --repair deletes orphaned embeddings (range-based for chunks)\n- [ ] --repair clears stale metadata (document_hash != content_hash at chunk_index=0)\n- [ ] --repair uses FTS `rebuild` (not `optimize`) for correct-by-construction repair\n- [ ] --repair requires --check (Clap `requires` attribute)\n- [ ] Human output: formatted with aligned columns\n- [ ] JSON output: `{\"ok\": true, \"data\": stats}`\n- [ ] `cargo build` succeeds\n\n## Files\n- `src/cli/commands/stats.rs` — new file\n- `src/cli/commands/mod.rs` — add `pub mod stats;`\n- `src/cli/mod.rs` — add StatsArgs, wire up stats subcommand\n- `src/main.rs` — add stats command handler\n\n## TDD Loop\nRED: Integration tests:\n- `test_stats_empty_db` — all counts 0, coverage 0%\n- `test_stats_with_documents` — correct counts by type\n- `test_integrity_check_healthy` — ok=true when consistent\n- `test_integrity_check_fts_mismatch` — detects FTS/doc count divergence\n- `test_integrity_check_orphan_embeddings` — detects orphaned rowids\n- `test_repair_rebuilds_fts` — FTS count matches after repair\n- `test_repair_cleans_orphans` — orphaned embeddings deleted\n- `test_repair_clears_stale` — stale metadata cleared (doc_hash mismatch)\nGREEN: Implement stats, integrity check, repair\nVERIFY: `cargo build && cargo test stats`\n\n## Edge Cases\n- Empty database: all counts 0, coverage 0%, no integrity issues\n- Gate A only (no embeddings table): skip embedding stats gracefully\n- --repair on healthy DB: no-op, reports \"no issues found\" / zero counts\n- FTS rebuild on large DB: may be slow\n- --repair without --check: Clap rejects (requires attribute enforces dependency)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-30T15:26:50.232629Z","created_by":"tayloreernisse","updated_at":"2026-01-30T17:54:31.065586Z","closed_at":"2026-01-30T17:54:31.065501Z","close_reason":"Implemented stats CLI with document counts by type, embedding coverage, FTS index count, queue stats, --check integrity (FTS mismatch, orphan embeddings, stale metadata), --repair (rebuild FTS, delete orphans, clear stale). Human + JSON output. Builds clean.","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-pr1","depends_on_id":"bd-3qs","type":"blocks","created_at":"2026-01-30T15:29:24.806108Z","created_by":"tayloreernisse"}]}
 {"id":"bd-ser","title":"Implement MR ingestion module","description":"## Background\nMR ingestion module with cursor-based sync. Follows the same pattern as issue ingestion from CP1. Discussion sync eligibility is determined via DB query AFTER ingestion (not in-memory collection) to avoid memory growth on large projects.\n\n## Approach\nCreate `src/ingestion/merge_requests.rs` with:\n1. `IngestMergeRequestsResult` - Aggregated stats\n2. `ingest_merge_requests()` - Main ingestion function\n3. `upsert_merge_request()` - Single MR upsert\n4. Helper functions for labels, assignees, reviewers, cursor management\n\n## Files\n- `src/ingestion/merge_requests.rs` - New module\n- `src/ingestion/mod.rs` - Export new module\n- `tests/mr_ingestion_tests.rs` - Integration tests\n\n## Acceptance Criteria\n- [ ] `IngestMergeRequestsResult` has: fetched, upserted, labels_created, assignees_linked, reviewers_linked\n- [ ] `ingest_merge_requests()` returns `Result<IngestMergeRequestsResult>`\n- [ ] Page-boundary cursor updates (not item-count modulo)\n- [ ] Tuple-based cursor filtering: `(updated_at, gitlab_id)`\n- [ ] Transaction per MR for atomicity\n- [ ] Raw payload stored for each MR\n- [ ] Labels: clear-and-relink pattern (removes stale)\n- [ ] Assignees: clear-and-relink pattern\n- [ ] Reviewers: clear-and-relink pattern\n- [ ] `reset_discussion_watermarks()` for --full sync\n- [ ] `cargo test mr_ingestion` passes\n\n## TDD Loop\nRED: `cargo test ingest_mr` -> module not found\nGREEN: Add ingestion module with full logic\nVERIFY: `cargo test mr_ingestion`\n\n## Main Function Signature\n```rust\npub async fn ingest_merge_requests(\n    conn: &Connection,\n    client: &GitLabClient,\n    config: &Config,\n    project_id: i64,           // Local DB project ID\n    gitlab_project_id: i64,    // GitLab project ID\n    full_sync: bool,           // Reset cursor if true\n) -> Result<IngestMergeRequestsResult>\n```\n\n## Ingestion Loop (page-based)\n```rust\nlet mut page = 1u32;\nloop {\n    let page_result = client.fetch_merge_requests_page(...).await?;\n    \n    for mr in &page_result.items {\n        // Tuple cursor filtering\n        if let (Some(cursor_ts), Some(cursor_id)) = (cursor_updated_at, cursor_gitlab_id) {\n            if mr_updated_at < cursor_ts { continue; }\n            if mr_updated_at == cursor_ts && mr.id <= cursor_id { continue; }\n        }\n        \n        // Begin transaction\n        let tx = conn.unchecked_transaction()?;\n        \n        // Store raw payload\n        let payload_id = store_payload(&tx, ...)?;\n        \n        // Transform and upsert\n        let transformed = transform_merge_request(&mr, project_id)?;\n        let upsert_result = upsert_merge_request(&tx, &transformed.merge_request, payload_id)?;\n        \n        // Clear-and-relink labels\n        clear_mr_labels(&tx, local_mr_id)?;\n        for label in &labels { ... }\n        \n        // Clear-and-relink assignees\n        clear_mr_assignees(&tx, local_mr_id)?;\n        for username in &transformed.assignee_usernames { ... }\n        \n        // Clear-and-relink reviewers\n        clear_mr_reviewers(&tx, local_mr_id)?;\n        for username in &transformed.reviewer_usernames { ... }\n        \n        tx.commit()?;\n        \n        // Track for cursor\n        last_updated_at = Some(mr_updated_at);\n        last_gitlab_id = Some(mr.id);\n    }\n    \n    // Page-boundary cursor flush\n    if let (Some(updated_at), Some(gitlab_id)) = (last_updated_at, last_gitlab_id) {\n        update_cursor(conn, project_id, \"merge_requests\", updated_at, gitlab_id)?;\n    }\n    \n    if page_result.is_last_page { break; }\n    page = page_result.next_page.unwrap_or(page + 1);\n}\n```\n\n## Full Sync Watermark Reset\n```rust\nfn reset_discussion_watermarks(conn: &Connection, project_id: i64) -> Result<()> {\n    conn.execute(\n        \"UPDATE merge_requests\n         SET discussions_synced_for_updated_at = NULL,\n             discussions_sync_attempts = 0,\n             discussions_sync_last_error = NULL\n         WHERE project_id = ?\",\n        [project_id],\n    )?;\n    Ok(())\n}\n```\n\n## DB Helper Functions\n- `get_cursor(conn, project_id) -> (Option<i64>, Option<i64>)` - Get (updated_at, gitlab_id)\n- `update_cursor(conn, project_id, resource_type, updated_at, gitlab_id)`\n- `reset_cursor(conn, project_id, resource_type)`\n- `upsert_merge_request(conn, mr, payload_id) -> Result<UpsertMergeRequestResult>`\n- `clear_mr_labels(conn, mr_id)`\n- `link_mr_label(conn, mr_id, label_id)`\n- `clear_mr_assignees(conn, mr_id)`\n- `upsert_mr_assignee(conn, mr_id, username)`\n- `clear_mr_reviewers(conn, mr_id)`\n- `upsert_mr_reviewer(conn, mr_id, username)`\n\n## Edge Cases\n- Cursor rewind may cause refetch of already-seen MRs (tuple filtering handles this)\n- Large projects: 10k+ MRs - page-based cursor prevents massive refetch on crash\n- Labels/assignees/reviewers may change - clear-and-relink ensures correctness","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-26T22:06:41.967459Z","created_by":"tayloreernisse","updated_at":"2026-01-27T00:15:24.526208Z","closed_at":"2026-01-27T00:15:24.526142Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-ser","depends_on_id":"bd-34o","type":"blocks","created_at":"2026-01-26T22:08:54.519486Z","created_by":"tayloreernisse"},{"issue_id":"bd-ser","depends_on_id":"bd-3ir","type":"blocks","created_at":"2026-01-26T22:08:54.440174Z","created_by":"tayloreernisse"},{"issue_id":"bd-ser","depends_on_id":"bd-iba","type":"blocks","created_at":"2026-01-26T22:08:54.593550Z","created_by":"tayloreernisse"}]}
+{"id":"bd-sqw","title":"Add Resource Events API endpoints to GitLab client","description":"## Background\nNeed paginated fetching of state/label/milestone events per entity from GitLab Resource Events APIs. The existing client uses reqwest with rate limiting and has stream_issues/stream_merge_requests patterns for paginated endpoints. However, resource events are per-entity (not project-wide), so they should return Vec rather than use streaming.\n\nExisting pagination pattern in client.rs: follow Link headers with per_page=100.\n\n## Approach\nAdd to src/gitlab/client.rs a generic helper and 6 endpoint methods:\n\n1. Generic paginated fetch helper (if not already present):\n```rust\nasync fn fetch_all_pages<T: DeserializeOwned>(&self, url: &str) -> Result<Vec<T>> {\n    let mut results = Vec::new();\n    let mut next_url = Some(url.to_string());\n    while let Some(current_url) = next_url {\n        self.rate_limiter.lock().unwrap().wait();\n        let resp = self.client.get(&current_url)\n            .header(\"PRIVATE-TOKEN\", &self.token)\n            .query(&[(\"per_page\", \"100\")])\n            .send().await?;\n        // ... parse Link header for next page\n        let page: Vec<T> = resp.json().await?;\n        results.extend(page);\n        next_url = parse_next_link(&resp_headers);\n    }\n    Ok(results)\n}\n```\n\n2. Six endpoint methods:\n```rust\npub async fn fetch_issue_state_events(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabStateEvent>>\npub async fn fetch_issue_label_events(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabLabelEvent>>\npub async fn fetch_issue_milestone_events(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabMilestoneEvent>>\npub async fn fetch_mr_state_events(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabStateEvent>>\npub async fn fetch_mr_label_events(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabLabelEvent>>\npub async fn fetch_mr_milestone_events(&self, project_id: i64, iid: i64) -> Result<Vec<GitLabMilestoneEvent>>\n```\n\nURL patterns:\n- Issues: `/api/v4/projects/{project_id}/issues/{iid}/resource_{type}_events`\n- MRs: `/api/v4/projects/{project_id}/merge_requests/{iid}/resource_{type}_events`\n\n3. Consider a convenience method that fetches all 3 event types for an entity in one call:\n```rust\npub async fn fetch_all_resource_events(&self, project_id: i64, entity_type: &str, iid: i64) \n    -> Result<(Vec<GitLabStateEvent>, Vec<GitLabLabelEvent>, Vec<GitLabMilestoneEvent>)>\n```\n\n## Acceptance Criteria\n- [ ] All 6 endpoints construct correct URLs\n- [ ] Pagination follows Link headers (handles entities with >100 events)\n- [ ] Rate limiter respected for each page request\n- [ ] 404 returns GitLabNotFound error (entity may have been deleted)\n- [ ] Network errors wrapped in GitLabNetworkError\n- [ ] Types from bd-2fm used for deserialization\n\n## Files\n- src/gitlab/client.rs (add methods + optionally generic helper)\n\n## TDD Loop\nRED: Add to tests/gitlab_client_tests.rs (or new file):\n- `test_fetch_issue_state_events_url` - verify URL construction (mock or inspect)\n- `test_fetch_mr_label_events_url` - verify URL construction\n- Note: Full integration tests require a mock HTTP server (mockito or wiremock). If the project doesn't already have one, write URL-construction unit tests only.\n\nGREEN: Implement the 6 methods using the generic helper\n\nVERIFY: `cargo test gitlab_client -- --nocapture && cargo build`\n\n## Edge Cases\n- project_id here is the GitLab project ID (not local DB id) — callers must pass gitlab_project_id\n- Empty results (new entity with no events) should return Ok(Vec::new()), not error\n- GitLab returns 403 for projects where Resource Events API is disabled — map to appropriate error\n- Very old entities may have thousands of events — pagination is essential\n- Rate limiter must be called per-page, not per-entity","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:24.137296Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:41:52.108185Z","compaction_level":0,"original_size":0,"labels":["api","gate-1","phase-b"],"dependencies":[{"issue_id":"bd-sqw","depends_on_id":"bd-2fm","type":"blocks","created_at":"2026-02-02T21:32:06.101374Z","created_by":"tayloreernisse"},{"issue_id":"bd-sqw","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:24.138647Z","created_by":"tayloreernisse"}]}
+{"id":"bd-tir","title":"Implement generic dependent fetch queue (enqueue + drain)","description":"## Background\nThe pending_dependent_fetches table (migration 011) provides a generic job queue for all dependent resource fetches across Gates 1, 2, and 4. This module implements the queue operations: enqueue, claim, complete, fail, and stale lock reclamation. It generalizes the existing discussion_queue.rs pattern.\n\n## Approach\nCreate src/core/dependent_queue.rs with:\n\n```rust\nuse rusqlite::Connection;\nuse super::error::Result;\n\n/// A pending job from the dependent fetch queue.\npub struct PendingJob {\n    pub id: i64,\n    pub project_id: i64,\n    pub entity_type: String,      // \"issue\" | \"merge_request\"\n    pub entity_iid: i64,\n    pub entity_local_id: i64,\n    pub job_type: String,         // \"resource_events\" | \"mr_closes_issues\" | \"mr_diffs\"\n    pub payload_json: Option<String>,\n    pub attempts: i32,\n}\n\n/// Enqueue a dependent fetch job. Idempotent via UNIQUE constraint (INSERT OR IGNORE).\npub fn enqueue_job(\n    conn: &Connection,\n    project_id: i64,\n    entity_type: &str,\n    entity_iid: i64,\n    entity_local_id: i64,\n    job_type: &str,\n    payload_json: Option<&str>,\n) -> Result<bool>  // returns true if actually inserted (not deduped)\n\n/// Claim a batch of jobs for processing. Atomically sets locked_at.\n/// Only claims jobs where locked_at IS NULL AND (next_retry_at IS NULL OR next_retry_at <= now).\npub fn claim_jobs(\n    conn: &Connection,\n    job_type: &str,\n    batch_size: usize,\n) -> Result<Vec<PendingJob>>\n\n/// Mark a job as complete (DELETE the row).\npub fn complete_job(conn: &Connection, job_id: i64) -> Result<()>\n\n/// Mark a job as failed. Increment attempts, set next_retry_at with exponential backoff, clear locked_at.\n/// Backoff: 30s * 2^(attempts-1), capped at 480s.\npub fn fail_job(conn: &Connection, job_id: i64, error: &str) -> Result<()>\n\n/// Reclaim stale locks (locked_at older than threshold).\n/// Returns count of reclaimed jobs.\npub fn reclaim_stale_locks(conn: &Connection, stale_threshold_minutes: u32) -> Result<usize>\n\n/// Count pending jobs by job_type (for stats/progress).\npub fn count_pending_jobs(conn: &Connection) -> Result<HashMap<String, usize>>\n```\n\nRegister in src/core/mod.rs: `pub mod dependent_queue;`\n\n**Key implementation details:**\n- claim_jobs uses a two-step approach: SELECT ids WHERE available, then UPDATE SET locked_at for those ids. Use a single transaction.\n- enqueued_at = current time in ms epoch UTC\n- locked_at = current time in ms epoch UTC when claimed\n- Backoff formula: next_retry_at = now + min(30_000 * 2^(attempts-1), 480_000) ms\n\n## Acceptance Criteria\n- [ ] enqueue_job is idempotent (INSERT OR IGNORE on UNIQUE constraint)\n- [ ] enqueue_job returns true on insert, false on dedup\n- [ ] claim_jobs only claims unlocked, non-retrying jobs\n- [ ] claim_jobs respects batch_size limit\n- [ ] complete_job DELETEs the row\n- [ ] fail_job increments attempts, sets next_retry_at, clears locked_at, records last_error\n- [ ] Backoff: 30s, 60s, 120s, 240s, 480s (capped)\n- [ ] reclaim_stale_locks clears locked_at for jobs older than threshold\n- [ ] count_pending_jobs returns accurate counts by job_type\n\n## Files\n- src/core/dependent_queue.rs (new)\n- src/core/mod.rs (add `pub mod dependent_queue;`)\n\n## TDD Loop\nRED: tests/dependent_queue_tests.rs (new):\n- `test_enqueue_job_basic` - enqueue a job, verify it exists\n- `test_enqueue_job_idempotent` - enqueue same job twice, verify single row\n- `test_claim_jobs_batch` - enqueue 5, claim 3, verify 3 returned and locked\n- `test_claim_jobs_skips_locked` - lock a job, claim again, verify it's skipped\n- `test_claim_jobs_respects_retry_at` - set next_retry_at in future, verify skipped\n- `test_claim_jobs_includes_retryable` - set next_retry_at in past, verify claimed\n- `test_complete_job_deletes` - complete a job, verify gone\n- `test_fail_job_backoff` - fail 3 times, verify exponential next_retry_at values\n- `test_reclaim_stale_locks` - set old locked_at, reclaim, verify cleared\n\nSetup: create_test_db() with migrations 001-011, seed project + issue.\n\nGREEN: Implement all functions\n\nVERIFY: `cargo test dependent_queue -- --nocapture`\n\n## Edge Cases\n- claim_jobs with batch_size=0 should return empty vec (not error)\n- enqueue_job with invalid job_type will be rejected by CHECK constraint — map rusqlite error to LoreError\n- fail_job on a non-existent job_id should be a no-op (job may have been completed by another path)\n- reclaim_stale_locks with 0 threshold would reclaim everything — ensure threshold is reasonable (minimum 1 min)\n- Timestamps must use consistent ms epoch UTC (not seconds)","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:31:57.290181Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:42:39.331604Z","compaction_level":0,"original_size":0,"labels":["gate-1","phase-b","queue"],"dependencies":[{"issue_id":"bd-tir","depends_on_id":"bd-2zl","type":"parent-child","created_at":"2026-02-02T21:31:57.291894Z","created_by":"tayloreernisse"},{"issue_id":"bd-tir","depends_on_id":"bd-hu3","type":"blocks","created_at":"2026-02-02T21:31:57.292472Z","created_by":"tayloreernisse"}]}
 {"id":"bd-v6i","title":"[CP1] gi ingest --type=issues command","description":"## Background\n\nThe `gi ingest --type=issues` command is the main entry point for issue ingestion. It acquires a single-flight lock, calls the orchestrator for each configured project, and outputs progress/summary to the user.\n\n## Approach\n\n### Module: src/cli/commands/ingest.rs\n\n### Clap Definition\n\n```rust\n#[derive(Args)]\npub struct IngestArgs {\n    /// Resource type to ingest\n    #[arg(long, value_parser = [\"issues\", \"merge_requests\"])]\n    pub r#type: String,\n\n    /// Filter to single project\n    #[arg(long)]\n    pub project: Option<String>,\n\n    /// Override stale sync lock\n    #[arg(long)]\n    pub force: bool,\n}\n```\n\n### Handler Function\n\n```rust\npub async fn handle_ingest(args: IngestArgs, config: &Config) -> Result<()>\n```\n\n### Logic\n\n1. **Acquire single-flight lock**: `acquire_sync_lock(conn, args.force)?`\n2. **Get projects to sync**:\n   - If `args.project` specified, filter to that one\n   - Otherwise, get all configured projects from DB\n3. **For each project**:\n   - Print \"Ingesting issues for {project_path}...\"\n   - Call `ingest_project_issues(conn, client, config, project_id, gitlab_project_id)`\n   - Print \"{N} issues fetched, {M} new labels\"\n4. **Print discussion sync summary**:\n   - \"Fetching discussions ({N} issues with updates)...\"\n   - \"{N} discussions, {M} notes (excluding {K} system notes)\"\n   - \"Skipped discussion sync for {N} unchanged issues.\"\n5. **Release lock**: Lock auto-released when handler returns\n\n### Output Format (matches PRD)\n\n```\nIngesting issues...\n\n  group/project-one: 1,234 issues fetched, 45 new labels\n\nFetching discussions (312 issues with updates)...\n\n  group/project-one: 312 issues → 1,234 discussions, 5,678 notes\n\nTotal: 1,234 issues, 1,234 discussions, 5,678 notes (excluding 1,234 system notes)\nSkipped discussion sync for 922 unchanged issues.\n```\n\n## Acceptance Criteria\n\n- [ ] Clap args parse --type, --project, --force correctly\n- [ ] Single-flight lock acquired before sync starts\n- [ ] Lock error message is clear if concurrent run attempted\n- [ ] Progress output shows per-project counts\n- [ ] Summary includes unchanged issues skipped count\n- [ ] --force flag allows overriding stale lock\n\n## Files\n\n- src/cli/commands/mod.rs (add `pub mod ingest;`)\n- src/cli/commands/ingest.rs (create)\n- src/cli/mod.rs (add Ingest variant to Commands enum)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/cli_ingest_tests.rs\n#[tokio::test] async fn ingest_issues_acquires_lock()\n#[tokio::test] async fn ingest_issues_fails_on_concurrent_run()\n#[tokio::test] async fn ingest_issues_respects_project_filter()\n#[tokio::test] async fn ingest_issues_force_overrides_stale_lock()\n```\n\nGREEN: Implement handler with lock and orchestrator calls\n\nVERIFY: `cargo test cli_ingest`\n\n## Edge Cases\n\n- No projects configured - return early with helpful message\n- Project filter matches nothing - error with \"project not found\"\n- Lock already held - clear error \"Sync already in progress\"\n- Ctrl-C during sync - lock should be released (via Drop or SIGINT handler)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.312565Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:56:44.090142Z","closed_at":"2026-01-25T22:56:44.090086Z","close_reason":"done","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-v6i","depends_on_id":"bd-ozy","type":"blocks","created_at":"2026-01-25T17:04:05.629772Z","created_by":"tayloreernisse"}]}
 {"id":"bd-xhz","title":"[CP1] GitLab client pagination methods","description":"## Background\n\nGitLab pagination methods enable fetching large result sets (issues, discussions) as async streams. The client uses `x-next-page` headers to determine continuation and applies cursor rewind for tuple-based incremental sync.\n\n## Approach\n\nAdd pagination methods to GitLabClient using `async-stream` crate:\n\n### Methods to Add\n\n```rust\nimpl GitLabClient {\n    /// Paginate through issues for a project.\n    pub fn paginate_issues(\n        &self,\n        gitlab_project_id: i64,\n        updated_after: Option<i64>,      // ms epoch cursor\n        cursor_rewind_seconds: u32,\n    ) -> Pin<Box<dyn Stream<Item = Result<GitLabIssue>> + Send + '_>>\n\n    /// Paginate through discussions for an issue.\n    pub fn paginate_issue_discussions(\n        &self,\n        gitlab_project_id: i64,\n        issue_iid: i64,\n    ) -> Pin<Box<dyn Stream<Item = Result<GitLabDiscussion>> + Send + '_>>\n\n    /// Make request and return response with headers for pagination.\n    async fn request_with_headers<T: DeserializeOwned>(\n        &self,\n        path: &str,\n        params: &[(&str, String)],\n    ) -> Result<(T, HeaderMap)>\n}\n```\n\n### Pagination Logic\n\n1. Start at page 1, per_page=100\n2. For issues: add scope=all, state=all, order_by=updated_at, sort=asc\n3. Apply cursor rewind: `updated_after = cursor - rewind_seconds` (clamped to 0)\n4. Yield each item from response\n5. Check `x-next-page` header for continuation\n6. Stop when header is empty/absent OR response is empty\n\n### Cursor Rewind\n\n```rust\nif let Some(ts) = updated_after {\n    let rewind_ms = (cursor_rewind_seconds as i64) * 1000;\n    let rewound = (ts - rewind_ms).max(0);  // Clamp to avoid underflow\n    // Convert to ISO 8601 for updated_after param\n}\n```\n\n## Acceptance Criteria\n\n- [ ] `paginate_issues` returns Stream of GitLabIssue\n- [ ] `paginate_issues` adds scope=all, state=all, order_by=updated_at, sort=asc\n- [ ] `paginate_issues` applies cursor rewind with max(0) clamping\n- [ ] `paginate_issue_discussions` returns Stream of GitLabDiscussion\n- [ ] Both methods follow x-next-page header until empty\n- [ ] Both methods stop on empty response (fallback)\n- [ ] `request_with_headers` returns (T, HeaderMap) tuple\n\n## Files\n\n- src/gitlab/client.rs (edit - add methods)\n\n## TDD Loop\n\nRED:\n```rust\n// tests/pagination_tests.rs\n#[tokio::test] async fn fetches_all_pages_when_multiple_exist()\n#[tokio::test] async fn respects_per_page_parameter()\n#[tokio::test] async fn follows_x_next_page_header_until_empty()\n#[tokio::test] async fn falls_back_to_empty_page_stop_if_headers_missing()\n#[tokio::test] async fn applies_cursor_rewind_for_tuple_semantics()\n#[tokio::test] async fn clamps_negative_rewind_to_zero()\n```\n\nGREEN: Implement pagination methods with async-stream\n\nVERIFY: `cargo test pagination`\n\n## Edge Cases\n\n- cursor_updated_at near zero - rewind must not underflow (use max(0))\n- GitLab returns empty x-next-page - treat as end of pages\n- GitLab omits pagination headers entirely - use empty response as stop condition\n- DateTime conversion fails - omit updated_after and fetch all (safe fallback)","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.222168Z","created_by":"tayloreernisse","updated_at":"2026-01-25T22:28:39.192876Z","closed_at":"2026-01-25T22:28:39.192815Z","close_reason":"Implemented paginate_issues and paginate_issue_discussions with async-stream, cursor rewind with max(0) clamping, x-next-page header following, 4 unit tests passing","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-xhz","depends_on_id":"bd-1np","type":"blocks","created_at":"2026-01-25T17:04:05.398212Z","created_by":"tayloreernisse"},{"issue_id":"bd-xhz","depends_on_id":"bd-2ys","type":"blocks","created_at":"2026-01-25T17:04:05.371440Z","created_by":"tayloreernisse"}]}
 {"id":"bd-ymd","title":"[CP1] Final validation - Gate A through D","description":"Run all tests and verify all internal gates pass.\n\n## Gate A: Issues Only (Must Pass First)\n- [ ] gi ingest --type=issues fetches all issues from configured projects\n- [ ] Issues stored with correct schema, including last_seen_at\n- [ ] Cursor-based sync is resumable (re-run fetches only new/updated)\n- [ ] Incremental cursor updates every 100 issues\n- [ ] Raw payloads stored for each issue\n- [ ] gi list issues and gi count issues work\n\n## Gate B: Labels Correct (Must Pass)\n- [ ] Labels extracted and stored (name-only)\n- [ ] Label links created correctly\n- [ ] Stale label links removed on re-sync (verified with test)\n- [ ] Label count per issue matches GitLab\n\n## Gate C: Dependent Discussion Sync (Must Pass)\n- [ ] Discussions fetched for issues with updated_at advancement\n- [ ] Notes stored with is_system flag correctly set\n- [ ] Raw payloads stored for discussions and notes\n- [ ] discussions_synced_for_updated_at watermark updated after sync\n- [ ] Unchanged issues skip discussion refetch (verified with test)\n- [ ] Bounded concurrency (dependent_concurrency respected)\n\n## Gate D: Resumability Proof (Must Pass)\n- [ ] Kill mid-run, rerun; bounded redo (cursor progress preserved)\n- [ ] No redundant discussion refetch after crash recovery\n- [ ] Single-flight lock prevents concurrent runs\n\n## Final Gate (Must Pass)\n- [ ] All unit tests pass (cargo test)\n- [ ] All integration tests pass (mocked with wiremock)\n- [ ] cargo clippy passes with no warnings\n- [ ] cargo fmt --check passes\n- [ ] Compiles with --release\n\n## Validation Commands\ncargo test\ncargo clippy -- -D warnings\ncargo fmt --check\ncargo build --release\n\nFiles: All CP1 files\nDone when: All gate criteria pass","status":"tombstone","priority":2,"issue_type":"task","created_at":"2026-01-25T16:59:26.795633Z","created_by":"tayloreernisse","updated_at":"2026-01-25T17:02:02.132613Z","deleted_at":"2026-01-25T17:02:02.132608Z","deleted_by":"tayloreernisse","delete_reason":"recreating with correct deps","original_type":"task","compaction_level":0,"original_size":0}
+{"id":"bd-ypa","title":"Implement timeline expand phase: BFS cross-reference expansion","description":"## Background\nStep 3 of timeline query: follow entity_references to discover related entities not matched by keywords. Uses BFS traversal over the entity_references table with configurable depth.\n\n## Approach\nAdd to src/core/timeline.rs:\n\n```rust\n/// Expand seed entities by following entity_references via BFS.\n/// Returns discovered entities with provenance tracking.\npub fn expand_via_references(\n    conn: &Connection,\n    seeds: &[SeedEntity],\n    depth: usize,             // --depth flag (default 1, 0 disables)\n    expand_mentions: bool,    // --expand-mentions flag (default false)\n    max_entities: usize,      // cap total expanded entities (default 100)\n) -> Result<ExpansionResult>\n\npub struct ExpansionResult {\n    pub expanded: Vec<ExpandedEntity>,\n    pub unresolved: Vec<UnresolvedRef>,\n}\n\npub struct ExpandedEntity {\n    pub entity_type: String,\n    pub local_id: i64,\n    pub iid: i64,\n    pub project_id: i64,\n    pub project_path: String,\n    pub depth: usize,\n    pub via_entity_type: String,\n    pub via_entity_iid: i64,\n    pub via_project: String,\n    pub via_reference_type: String,\n    pub via_source_method: String,\n}\n```\n\nBFS algorithm:\n1. Initialize frontier = set of seed entity (type, local_id) pairs\n2. Initialize visited = frontier.clone()\n3. For each depth level 1..=depth:\n   a. Query entity_references WHERE source OR target is in frontier\n   b. Filter edge types: always include 'closes' + 'related'; include 'mentioned' only if expand_mentions\n   c. For each ref, the \"other side\" (not in visited) is a candidate\n   d. Resolve candidate to entity details (iid, project_path, etc.)\n   e. Skip if already visited\n   f. Skip unresolved refs (target_entity_id IS NULL) — collect in unresolved list\n   g. Add to expanded list with provenance\n   h. Add to visited + next frontier\n   i. Stop if total expanded >= max_entities\n4. Return expanded entities + unresolved refs\n\nSQL for finding adjacent entities:\n```sql\n-- Outgoing edges (from seed entities)\nSELECT target_entity_type, target_entity_id, target_project_path, target_entity_iid,\n       reference_type, source_method\nFROM entity_references\nWHERE source_entity_type = ?1 AND source_entity_id = ?2\n  AND reference_type IN ('closes', 'related' [, 'mentioned'])\n\n-- Incoming edges (to seed entities)  \nSELECT source_entity_type, source_entity_id, NULL, NULL,\n       reference_type, source_method\nFROM entity_references\nWHERE target_entity_type = ?1 AND target_entity_id = ?2\n  AND reference_type IN ('closes', 'related' [, 'mentioned'])\n```\n\n## Acceptance Criteria\n- [ ] BFS correctly traverses to depth N\n- [ ] depth=0 returns empty expansion (no traversal)\n- [ ] Default: only 'closes' + 'related' edges followed\n- [ ] --expand-mentions adds 'mentioned' edges\n- [ ] Provenance tracked: which seed entity and edge type led to discovery\n- [ ] Unresolved refs (target_entity_id NULL) collected but not traversed\n- [ ] max_entities cap prevents explosion\n- [ ] No cycles (visited set prevents re-expansion)\n\n## Files\n- src/core/timeline.rs (add expand_via_references)\n\n## TDD Loop\nRED: tests/timeline_expand_tests.rs:\n- `test_expand_depth_1_follows_closes` - seed issue, MR closes it → MR expanded\n- `test_expand_depth_0_returns_empty` - no expansion\n- `test_expand_depth_2_two_hops` - issue → MR → issue chain\n- `test_expand_skips_mentioned_by_default` - 'mentioned' edge not followed\n- `test_expand_includes_mentioned_when_flag` - 'mentioned' edge followed with flag\n- `test_expand_collects_unresolved` - cross-project ref in unresolved list\n- `test_expand_caps_at_max_entities` - 200 reachable entities, cap at 100\n- `test_expand_no_cycles` - circular reference doesn't infinite loop\n\nSetup: create_test_db, seed entity_references with various edge types.\n\nGREEN: Implement BFS traversal\n\nVERIFY: `cargo test timeline_expand -- --nocapture`\n\n## Edge Cases\n- Bidirectional traversal: a 'closes' edge from MR→issue should be traversable from the issue side too (find MRs that close this issue)\n- Self-referential refs (entity references itself) — visited set handles this\n- Large fan-out on popular entities — max_entities cap is essential\n- Entity may have been deleted since ref was created — skip missing entities gracefully","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:33:08.659381Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:46:11.677952Z","compaction_level":0,"original_size":0,"labels":["gate-3","phase-b","query"],"dependencies":[{"issue_id":"bd-ypa","depends_on_id":"bd-32q","type":"blocks","created_at":"2026-02-02T21:33:37.448515Z","created_by":"tayloreernisse"},{"issue_id":"bd-ypa","depends_on_id":"bd-3ia","type":"blocks","created_at":"2026-02-02T21:33:37.528233Z","created_by":"tayloreernisse"},{"issue_id":"bd-ypa","depends_on_id":"bd-ike","type":"parent-child","created_at":"2026-02-02T21:33:08.661036Z","created_by":"tayloreernisse"}]}
 {"id":"bd-z0s","title":"[CP1] Final validation - Gate A through D","description":"Run all tests and verify all internal gates pass.\n\n## Gate A: Issues Only (Must Pass First)\n- [ ] gi ingest --type=issues fetches all issues from configured projects\n- [ ] Issues stored with correct schema, including last_seen_at\n- [ ] Cursor-based sync is resumable (re-run fetches only new/updated)\n- [ ] Incremental cursor updates every 100 issues\n- [ ] Raw payloads stored for each issue\n- [ ] gi list issues and gi count issues work\n\n## Gate B: Labels Correct (Must Pass)\n- [ ] Labels extracted and stored (name-only)\n- [ ] Label links created correctly\n- [ ] **Stale label links removed on re-sync** (verified with test)\n- [ ] Label count per issue matches GitLab\n\n## Gate C: Dependent Discussion Sync (Must Pass)\n- [ ] Discussions fetched for issues with updated_at advancement\n- [ ] Notes stored with is_system flag correctly set\n- [ ] Raw payloads stored for discussions and notes\n- [ ] discussions_synced_for_updated_at watermark updated after sync\n- [ ] **Unchanged issues skip discussion refetch** (verified with test)\n- [ ] Bounded concurrency (dependent_concurrency respected)\n\n## Gate D: Resumability Proof (Must Pass)\n- [ ] Kill mid-run, rerun; bounded redo (cursor progress preserved)\n- [ ] No redundant discussion refetch after crash recovery\n- [ ] Single-flight lock prevents concurrent runs\n\n## Final Gate (Must Pass)\n- [ ] All unit tests pass (cargo test)\n- [ ] All integration tests pass (mocked with wiremock)\n- [ ] cargo clippy passes with no warnings\n- [ ] cargo fmt --check passes\n- [ ] Compiles with --release\n\n## Validation Commands\ncargo test\ncargo clippy -- -D warnings\ncargo fmt --check\ncargo build --release\n\n## Data Integrity Checks\n- SELECT COUNT(*) FROM issues matches GitLab issue count\n- Every issue has a raw_payloads row\n- Every discussion has a raw_payloads row\n- Labels in issue_labels junction all exist in labels table\n- Re-running gi ingest --type=issues fetches 0 new items\n- After removing a label in GitLab and re-syncing, the link is removed\n\nFiles: All CP1 files\nDone when: All gate criteria pass","status":"closed","priority":2,"issue_type":"task","created_at":"2026-01-25T17:02:38.459095Z","created_by":"tayloreernisse","updated_at":"2026-01-25T23:27:09.567537Z","closed_at":"2026-01-25T23:27:09.567478Z","close_reason":"All gates pass: 71 tests, clippy clean, fmt clean, release build successful","compaction_level":0,"original_size":0,"dependencies":[{"issue_id":"bd-z0s","depends_on_id":"bd-17v","type":"blocks","created_at":"2026-01-25T17:04:05.889114Z","created_by":"tayloreernisse"},{"issue_id":"bd-z0s","depends_on_id":"bd-2f0","type":"blocks","created_at":"2026-01-25T17:04:05.841210Z","created_by":"tayloreernisse"},{"issue_id":"bd-z0s","depends_on_id":"bd-39w","type":"blocks","created_at":"2026-01-25T17:04:05.913316Z","created_by":"tayloreernisse"},{"issue_id":"bd-z0s","depends_on_id":"bd-3n1","type":"blocks","created_at":"2026-01-25T17:04:05.817830Z","created_by":"tayloreernisse"},{"issue_id":"bd-z0s","depends_on_id":"bd-o7b","type":"blocks","created_at":"2026-01-25T17:04:05.864480Z","created_by":"tayloreernisse"},{"issue_id":"bd-z0s","depends_on_id":"bd-v6i","type":"blocks","created_at":"2026-01-25T17:04:05.794555Z","created_by":"tayloreernisse"}]}
+{"id":"bd-z94","title":"Implement 'lore file-history' command with human and robot output","description":"## Background\nlore file-history <path> shows MRs that touched a file, ordered by merge/creation date. Uses rename chain resolution (bd-1yx), file change data (bd-2yo), and cross-references (bd-3ia) for MR→issue linking.\n\n## Approach\nCreate src/cli/commands/file_history.rs:\n\n**1. Query logic:**\n```rust\npub fn run_file_history(\n    conn: &Connection,\n    path: &str,\n    project_id: Option<i64>,\n    follow_renames: bool,\n    merged_only: bool,\n    include_discussions: bool,\n    limit: usize,\n) -> Result<FileHistoryResult>\n\npub struct FileHistoryResult {\n    pub file_path: String,\n    pub rename_chain: Option<Vec<String>>,\n    pub mrs: Vec<FileHistoryMR>,\n}\n\npub struct FileHistoryMR {\n    pub iid: i64,\n    pub title: String,\n    pub state: String,\n    pub author: String,\n    pub change_type: String,\n    pub old_path: Option<String>,\n    pub merged_at: Option<i64>,\n    pub created_at: i64,\n    pub web_url: String,\n    pub merge_commit_sha: Option<String>,\n    pub discussion_count: usize,\n    pub file_discussion_count: usize,\n    pub closes_issues: Vec<LinkedIssue>,\n    pub discussions: Option<Vec<FileDiscussion>>,  // when --discussions flag\n}\n```\n\n**2. SQL query** (from spec §4.5):\n```sql\nSELECT mr.iid, mr.title, mr.state, mr.author_username,\n       mr.merged_at, mr.created_at, mr.web_url, mr.merge_commit_sha,\n       mfc.change_type, mfc.old_path,\n       (SELECT COUNT(*) FROM discussions d WHERE d.merge_request_id = mr.id) AS discussion_count,\n       (SELECT COUNT(*) FROM notes n JOIN discussions d ON n.discussion_id = d.id\n        WHERE d.merge_request_id = mr.id AND n.position_new_path = ?1) AS file_discussion_count\nFROM mr_file_changes mfc\nJOIN merge_requests mr ON mr.id = mfc.merge_request_id\nWHERE mfc.new_path IN (rename_chain_paths) OR mfc.old_path IN (rename_chain_paths)\nORDER BY COALESCE(mr.merged_at, mr.created_at) DESC\nLIMIT ?;\n```\n\n**3. Ambiguous project handling:**\nIf path exists in multiple projects and -p not specified, return Ambiguous error (exit code 18).\n\n**4. Human output:**\n```\nFile History: src/auth/oauth.rs\n(renamed from src/auth/handler.rs <- src/auth.rs)\n─────────────────────────────────────────────────\n!567  feat: add OAuth2 provider     MERGED  2024-03-25  @bob   modified\n      Closes #234: Migrate to OAuth2\n      12 discussions, 4 on this file\n!612  fix: token refresh race       MERGED  2024-04-10  @alice modified\n      Closes #299: OAuth2 login fails\n      5 discussions, 2 on this file\n```\n\n**5. Robot JSON:**\n```json\n{\n  \"ok\": true,\n  \"data\": {\n    \"file_path\": \"src/auth/oauth.rs\",\n    \"rename_chain\": [\"src/auth.rs\", \"src/auth/handler.rs\", \"src/auth/oauth.rs\"],\n    \"mrs\": [...]\n  },\n  \"meta\": { \"total_mrs\": 2 }\n}\n```\n\n## Acceptance Criteria\n- [ ] Correct MR results via rename chain resolution\n- [ ] MRs ordered by merge/creation date (newest first)\n- [ ] Discussion counts accurate (total + file-specific)\n- [ ] Rename chain displayed in human output\n- [ ] --discussions shows DiffNote snippets\n- [ ] --merged filters to merged MRs only\n- [ ] --no-follow-renames disables rename chain\n- [ ] -p required when path in multiple projects (Ambiguous error, exit 18)\n- [ ] Robot JSON includes rename_chain\n- [ ] Closes issues listed from entity_references\n- [ ] Graceful empty state: \"No MR data found. Run lore sync with fetchMrFileChanges: true\"\n\n## Files\n- src/cli/commands/file_history.rs (new)\n- src/cli/commands/mod.rs (add `pub mod file_history;`)\n- src/cli/mod.rs (add FileHistory variant + FileHistoryArgs)\n- src/main.rs (add handler)\n\n## TDD Loop\nRED: tests/file_history_command_tests.rs:\n- `test_file_history_basic` - seed file changes, verify MR results\n- `test_file_history_with_rename_chain` - renamed file, verify old MRs included\n- `test_file_history_merged_only` - mix of states, verify filter\n- `test_file_history_empty_state` - no data, verify graceful message\n- `test_file_history_ambiguous_project` - path in 2 projects without -p, verify error\n- `test_file_history_robot_json` - verify JSON structure\n\nGREEN: Implement command with query + rendering\n\nVERIFY: `cargo test file_history -- --nocapture`\n\n## Edge Cases\n- File with 100+ MRs: respect -n limit\n- Renamed file with no MRs at old path: rename chain still displayed but no extra MRs\n- File discussion count query: notes.position_new_path must match — some notes may have position_old_path instead\n- Deleted files: path still queryable if file_changes record exists","status":"open","priority":2,"issue_type":"task","created_at":"2026-02-02T21:34:09.027259Z","created_by":"tayloreernisse","updated_at":"2026-02-02T21:49:29.569291Z","compaction_level":0,"original_size":0,"labels":["cli","gate-4","phase-b"],"dependencies":[{"issue_id":"bd-z94","depends_on_id":"bd-14q","type":"parent-child","created_at":"2026-02-02T21:34:09.028633Z","created_by":"tayloreernisse"},{"issue_id":"bd-z94","depends_on_id":"bd-1yx","type":"blocks","created_at":"2026-02-02T21:34:16.784122Z","created_by":"tayloreernisse"},{"issue_id":"bd-z94","depends_on_id":"bd-2yo","type":"blocks","created_at":"2026-02-02T21:34:16.741201Z","created_by":"tayloreernisse"},{"issue_id":"bd-z94","depends_on_id":"bd-3ia","type":"blocks","created_at":"2026-02-02T21:34:16.824983Z","created_by":"tayloreernisse"}]}
diff --git a/.beads/last-touched b/.beads/last-touched
index 8530820..845f202 100644
--- a/.beads/last-touched
+++ b/.beads/last-touched
@@ -1 +1 @@
-bd-35o
+bd-1j1
diff --git a/.claude/agents/test-runner.md b/.claude/agents/test-runner.md
new file mode 100644
index 0000000..2a3a28c
--- /dev/null
+++ b/.claude/agents/test-runner.md
@@ -0,0 +1,59 @@
+---
+name: test-runner
+description: "Use this agent when unit tests need to be run and results analyzed. This includes after writing or modifying code, before committing changes, or when explicitly asked to verify test status.\\n\\nExamples:\\n\\n- User: \"Please refactor the parse_session function to handle edge cases\"\\n  Assistant: \"Here is the refactored function with edge case handling: ...\"\\n  [code changes applied]\\n  Since a significant piece of code was modified, use the Task tool to launch the test-runner agent to verify nothing is broken.\\n  Assistant: \"Now let me run the test suite to make sure everything still passes.\"\\n\\n- User: \"Do all tests pass?\"\\n  Assistant: \"Let me use the Task tool to launch the test-runner agent to check the current test status.\"\\n\\n- User: \"I just finished implementing the search feature\"\\n  Assistant: \"Let me use the Task tool to launch the test-runner agent to validate the implementation.\"\\n\\n- After any logical chunk of code is written or modified, proactively use the Task tool to launch the test-runner agent to run the tests before reporting completion to the user."
+tools: Bash
+model: haiku
+color: orange
+---
+
+You are an expert test execution and analysis engineer. Your sole responsibility is to run the project's unit test suite, interpret the results with precision, and deliver a clear, actionable summary.
+
+## Execution Protocol
+
+1. **Discover the test framework**: Examine the project structure to determine how tests are run:
+   - Look for `Cargo.toml` (Rust: `cargo test`)
+   - If unclear, check README or CLAUDE.md for test instructions
+
+2. **Run the tests**: Execute the appropriate test command. Capture full output including stdout and stderr. Do NOT run tests interactively or with watch mode. Use flags that produce verbose or detailed output when available (e.g., `cargo test -- --nocapture`, `jest --verbose`).
+
+3. **Analyze results**: Parse the test output carefully and categorize:
+   - Total tests run
+   - Tests passed
+   - Tests failed (with details)
+   - Tests skipped/ignored
+   - Compilation errors (if tests couldn't even run)
+
+4. **Report findings**:
+
+   **If ALL tests pass:**
+   Provide a concise success summary:
+   - Total test count and pass count
+   - Execution time if available
+   - Note any skipped/ignored tests and why (if apparent)
+   - A clear statement: "All tests passed."
+
+   **If ANY tests fail:**
+   Provide a detailed failure report:
+   - List each failing test by its full name/path
+   - Include the assertion error or panic message for each failure
+   - Include relevant expected vs actual values
+   - Note the file and line number where the failure occurred (if available)
+   - Group failures by module/file if there are many
+   - Suggest likely root causes when the error messages make it apparent
+   - Note if failures appear related (e.g., same underlying issue)
+
+   **If tests cannot run (compilation/setup error):**
+   - Report the exact error preventing test execution
+   - Identify the file and line causing the issue
+   - Distinguish between test code errors and source code errors
+
+## Rules
+
+- NEVER modify any source code or test code. You are read-only except for running the test command.
+- NEVER skip running tests and guess at results. Always execute the actual test command.
+- NEVER run the full application or any destructive commands. Only run test commands.
+- If the test suite is extremely large, run it fully anyway. Do not truncate or sample.
+- If multiple test targets exist (unit, integration, e2e), run unit tests only unless instructed otherwise.
+- Report raw numbers. Do not round or approximate test counts.
+- If tests produce warnings (not failures), mention them briefly but clearly separate them from failures.
+- Keep the summary structured and scannable. Use bullet points and clear headers.
diff --git a/api-review.html b/api-review.html
new file mode 100644
index 0000000..9ac6b9d
--- /dev/null
+++ b/api-review.html
@@ -0,0 +1,1654 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>Gitlore &harr; GitLab API Review</title>
+<style>
+  :root {
+    --bg: #0d1117;
+    --surface: #161b22;
+    --surface2: #1c2333;
+    --border: #30363d;
+    --text: #c9d1d9;
+    --text-muted: #8b949e;
+    --accent: #58a6ff;
+    --green: #3fb950;
+    --yellow: #d29922;
+    --red: #f85149;
+    --purple: #bc8cff;
+    --orange: #d18616;
+  }
+  * { margin: 0; padding: 0; box-sizing: border-box; }
+  body {
+    font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Helvetica, Arial, sans-serif;
+    background: var(--bg);
+    color: var(--text);
+    line-height: 1.6;
+  }
+  header {
+    background: var(--surface);
+    border-bottom: 1px solid var(--border);
+    padding: 20px 32px;
+    position: sticky;
+    top: 0;
+    z-index: 100;
+  }
+  header h1 { font-size: 20px; font-weight: 600; }
+  header p { color: var(--text-muted); font-size: 13px; margin-top: 4px; }
+
+  .tabs {
+    display: flex;
+    gap: 0;
+    border-bottom: 1px solid var(--border);
+    background: var(--surface);
+    padding: 0 32px;
+    overflow-x: auto;
+  }
+  .tab {
+    padding: 10px 16px;
+    cursor: pointer;
+    color: var(--text-muted);
+    font-size: 14px;
+    border-bottom: 2px solid transparent;
+    white-space: nowrap;
+    transition: color 0.15s, border-color 0.15s;
+    user-select: none;
+  }
+  .tab:hover { color: var(--text); }
+  .tab.active { color: var(--text); border-bottom-color: var(--accent); }
+
+  .content { max-width: 1200px; margin: 0 auto; padding: 24px 32px; }
+  .panel { display: none; }
+  .panel.active { display: block; }
+
+  h2 { font-size: 18px; margin-bottom: 16px; color: var(--text); }
+  h3 { font-size: 15px; margin: 20px 0 10px; color: var(--accent); }
+  h4 { font-size: 13px; margin: 14px 0 6px; color: var(--text-muted); text-transform: uppercase; letter-spacing: 0.5px; }
+
+  .card {
+    background: var(--surface);
+    border: 1px solid var(--border);
+    border-radius: 8px;
+    padding: 16px;
+    margin-bottom: 12px;
+  }
+  .card-title { font-weight: 600; font-size: 14px; margin-bottom: 8px; }
+
+  table {
+    width: 100%;
+    border-collapse: collapse;
+    font-size: 13px;
+    margin-bottom: 16px;
+  }
+  th {
+    text-align: left;
+    padding: 8px 12px;
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    color: var(--text-muted);
+    font-weight: 600;
+    font-size: 12px;
+    text-transform: uppercase;
+    letter-spacing: 0.3px;
+  }
+  td {
+    padding: 8px 12px;
+    border: 1px solid var(--border);
+    vertical-align: top;
+  }
+  tr:hover td { background: var(--surface2); }
+
+  code {
+    font-family: 'SF Mono', 'Fira Code', monospace;
+    font-size: 12px;
+    background: var(--surface2);
+    padding: 2px 6px;
+    border-radius: 4px;
+    color: var(--accent);
+  }
+  pre {
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 6px;
+    padding: 12px 16px;
+    overflow-x: auto;
+    font-size: 12px;
+    line-height: 1.5;
+    margin: 8px 0 16px;
+    font-family: 'SF Mono', 'Fira Code', monospace;
+    color: var(--text);
+  }
+
+  .badge {
+    display: inline-block;
+    padding: 2px 8px;
+    border-radius: 12px;
+    font-size: 11px;
+    font-weight: 600;
+    text-transform: uppercase;
+  }
+  .badge-get { background: #1a3a2a; color: var(--green); }
+  .badge-post { background: #2a2a1a; color: var(--yellow); }
+  .badge-put { background: #2a1a2a; color: var(--purple); }
+  .badge-delete { background: #2a1a1a; color: var(--red); }
+  .badge-used { background: #1a2a3a; color: var(--accent); }
+  .badge-unused { background: #1c1c1c; color: var(--text-muted); }
+  .badge-readonly { background: #1a3a2a; color: var(--green); }
+
+  .flow-diagram {
+    display: flex;
+    align-items: center;
+    gap: 8px;
+    flex-wrap: wrap;
+    margin: 12px 0;
+  }
+  .flow-box {
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 6px;
+    padding: 8px 14px;
+    font-size: 12px;
+    text-align: center;
+    min-width: 100px;
+  }
+  .flow-box.highlight { border-color: var(--accent); }
+  .flow-arrow { color: var(--text-muted); font-size: 18px; }
+
+  .grid-2 { display: grid; grid-template-columns: 1fr 1fr; gap: 12px; }
+  @media (max-width: 768px) { .grid-2 { grid-template-columns: 1fr; } }
+
+  .coverage-bar {
+    height: 8px;
+    background: var(--surface2);
+    border-radius: 4px;
+    margin: 6px 0;
+    overflow: hidden;
+  }
+  .coverage-fill {
+    height: 100%;
+    border-radius: 4px;
+    transition: width 0.3s;
+  }
+  .stat-row {
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    padding: 6px 0;
+    font-size: 13px;
+  }
+  .stat-label { color: var(--text-muted); }
+
+  .expandable-header {
+    cursor: pointer;
+    display: flex;
+    justify-content: space-between;
+    align-items: center;
+    padding: 10px 0;
+    border-bottom: 1px solid var(--border);
+    user-select: none;
+  }
+  .expandable-header:hover { color: var(--accent); }
+  .expandable-header .arrow { transition: transform 0.2s; }
+  .expandable-header.open .arrow { transform: rotate(90deg); }
+  .expandable-body { display: none; padding: 12px 0; }
+  .expandable-body.open { display: block; }
+
+  .search-box {
+    width: 100%;
+    padding: 8px 12px;
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 6px;
+    color: var(--text);
+    font-size: 13px;
+    margin-bottom: 16px;
+    outline: none;
+  }
+  .search-box:focus { border-color: var(--accent); }
+  .search-box::placeholder { color: var(--text-muted); }
+
+  .mapping-row {
+    display: grid;
+    grid-template-columns: 1fr 40px 1fr;
+    align-items: center;
+    gap: 8px;
+    padding: 8px 0;
+    border-bottom: 1px solid var(--border);
+    font-size: 13px;
+  }
+  .mapping-row:last-child { border-bottom: none; }
+  .mapping-connector { text-align: center; color: var(--accent); }
+
+  .legend { display: flex; gap: 16px; margin-bottom: 16px; flex-wrap: wrap; }
+  .legend-item { display: flex; align-items: center; gap: 6px; font-size: 12px; color: var(--text-muted); }
+  .legend-dot { width: 8px; height: 8px; border-radius: 50%; }
+
+  .f-stored td:first-child { border-left: 3px solid var(--green); }
+  .f-transient td:first-child { border-left: 3px solid var(--yellow); }
+  .f-ignored td:first-child { border-left: 3px solid var(--purple); }
+  .f-dropped td:first-child { border-left: 3px solid var(--red); }
+  .f-dropped td { color: var(--text-muted); }
+</style>
+</head>
+<body>
+
+<header>
+  <h1>Gitlore CLI &harr; GitLab API Review</h1>
+  <p>Interactive mapping of the gitlore local sync engine to the GitLab REST API v4</p>
+</header>
+
+<div class="tabs">
+  <div class="tab active" onclick="showTab('overview')">Overview</div>
+  <div class="tab" onclick="showTab('cli')">CLI Architecture</div>
+  <div class="tab" onclick="showTab('api-used')">API Endpoints Used</div>
+  <div class="tab" onclick="showTab('api-full')">Full GitLab API</div>
+  <div class="tab" onclick="showTab('mapping')">Mapping</div>
+  <div class="tab" onclick="showTab('data')">Data Model</div>
+  <div class="tab" onclick="showTab('pipeline')">Ingestion Pipeline</div>
+  <div class="tab" onclick="showTab('fields')">Field Coverage</div>
+  <div class="tab" onclick="showTab('efficiency')">Efficiency Analysis</div>
+</div>
+
+<div class="content">
+
+<!-- OVERVIEW -->
+<div id="overview" class="panel active">
+  <h2>What is Gitlore?</h2>
+  <div class="card" style="margin-bottom:24px">
+    <p>Gitlore (<code>lore</code>) is a <strong>read-only local sync engine</strong> for GitLab data. It fetches issues, merge requests, and discussions from the GitLab REST API v4 and stores them in a local SQLite database for fast offline querying.</p>
+    <p style="margin-top:8px;color:var(--text-muted)">~11,148 lines of Rust. Noun-first CLI. Robot mode for automation. Cursor-based incremental sync.</p>
+  </div>
+
+  <h3>API Surface Coverage</h3>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:12px">Gitlore uses a small, focused subset of the GitLab API. It is <strong>read-only</strong> &mdash; it never creates, updates, or deletes anything on GitLab.</p>
+
+  <div class="grid-2">
+    <div class="card">
+      <div class="card-title">Endpoints Used: 7</div>
+      <div class="stat-row"><span class="stat-label">Issues (list)</span><span class="badge badge-get">GET</span></div>
+      <div class="stat-row"><span class="stat-label">Merge Requests (list)</span><span class="badge badge-get">GET</span></div>
+      <div class="stat-row"><span class="stat-label">Issue Discussions (list)</span><span class="badge badge-get">GET</span></div>
+      <div class="stat-row"><span class="stat-label">MR Discussions (list)</span><span class="badge badge-get">GET</span></div>
+      <div class="stat-row"><span class="stat-label">Current User</span><span class="badge badge-get">GET</span></div>
+      <div class="stat-row"><span class="stat-label">Project Details</span><span class="badge badge-get">GET</span></div>
+      <div class="stat-row"><span class="stat-label">GitLab Version</span><span class="badge badge-get">GET</span></div>
+    </div>
+    <div class="card">
+      <div class="card-title">Coverage by Resource</div>
+      <div class="stat-row"><span>Issues API</span><span>2 of ~20 endpoints</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:10%;background:var(--green)"></div></div>
+      <div class="stat-row"><span>MR API</span><span>2 of ~30 endpoints</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:7%;background:var(--green)"></div></div>
+      <div class="stat-row"><span>Discussions API</span><span>2 of 30 endpoints</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:7%;background:var(--green)"></div></div>
+      <div class="stat-row"><span>Users API</span><span>1 of ~15 endpoints</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:7%;background:var(--green)"></div></div>
+      <div class="stat-row"><span>Projects API</span><span>1 of ~50 endpoints</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:2%;background:var(--green)"></div></div>
+    </div>
+  </div>
+
+  <h3>Data Flow</h3>
+  <div class="flow-diagram">
+    <div class="flow-box">GitLab API v4</div>
+    <div class="flow-arrow">&rarr;</div>
+    <div class="flow-box highlight">Rate Limiter<br><small>10 req/s + jitter</small></div>
+    <div class="flow-arrow">&rarr;</div>
+    <div class="flow-box">Async Streams<br><small>Paginated fetch</small></div>
+    <div class="flow-arrow">&rarr;</div>
+    <div class="flow-box">Transformers<br><small>Normalize data</small></div>
+    <div class="flow-arrow">&rarr;</div>
+    <div class="flow-box highlight">SQLite (WAL)<br><small>Local DB</small></div>
+  </div>
+
+  <h3>Key Design Decisions</h3>
+  <div class="grid-2">
+    <div class="card">
+      <div class="card-title">Read-only sync</div>
+      <p style="font-size:13px;color:var(--text-muted)">Only GET requests. Never mutates GitLab state. Safe to run repeatedly.</p>
+    </div>
+    <div class="card">
+      <div class="card-title">Cursor-based incremental</div>
+      <p style="font-size:13px;color:var(--text-muted)">Uses <code>updated_after</code> parameter to only fetch changed data. 2-second rewind overlap for safety.</p>
+    </div>
+    <div class="card">
+      <div class="card-title">Raw payload archival</div>
+      <p style="font-size:13px;color:var(--text-muted)">Stores original JSON responses with SHA-256 dedup and optional gzip compression.</p>
+    </div>
+    <div class="card">
+      <div class="card-title">Discussion full-refresh</div>
+      <p style="font-size:13px;color:var(--text-muted)">Discussions use DELETE+INSERT strategy per parent (no incremental). Parallel prefetch, serial write.</p>
+    </div>
+  </div>
+</div>
+
+<!-- CLI ARCHITECTURE -->
+<div id="cli" class="panel">
+  <h2>CLI Architecture</h2>
+
+  <h3>Command Structure (Noun-First)</h3>
+  <div class="card">
+    <pre>lore &lt;noun&gt; [verb/arg]    # Primary pattern
+lore issues              # List all issues
+lore issues 42           # Show issue #42
+lore mrs                 # List all merge requests
+lore mrs 17              # Show MR #17
+lore ingest issues       # Fetch issues from GitLab
+lore ingest mrs          # Fetch MRs from GitLab
+lore count issues        # Count local issues
+lore count discussions   # Count local discussions
+lore status              # Show sync state
+lore auth                # Verify GitLab auth
+lore doctor              # Health check
+lore init                # Initialize config + DB
+lore migrate             # Run DB migrations
+lore version             # Show version</pre>
+  </div>
+
+  <h3>Global Flags</h3>
+  <table>
+    <tr><th>Flag</th><th>Description</th></tr>
+    <tr><td><code>-c, --config</code></td><td>Path to config file</td></tr>
+    <tr><td><code>--robot</code></td><td>Machine-readable JSON output</td></tr>
+    <tr><td><code>-J, --json</code></td><td>JSON shorthand (same as --robot)</td></tr>
+  </table>
+
+  <h3>Robot Mode Detection</h3>
+  <div class="card">
+    <p style="font-size:13px">Three ways to activate:</p>
+    <pre>lore --robot list issues          # Explicit flag
+lore list issues | jq .           # Auto: stdout not a TTY
+LORE_ROBOT=1 lore list issues     # Environment variable</pre>
+    <p style="font-size:13px;color:var(--text-muted);margin-top:8px">Robot mode returns JSON: <code>{"ok":true,"data":{...},"meta":{...}}</code></p>
+    <p style="font-size:13px;color:var(--text-muted)">Errors go to stderr: <code>{"error":{"code":"...","message":"...","suggestion":"..."}}</code></p>
+  </div>
+
+  <h3>Exit Codes</h3>
+  <table>
+    <tr><th>Code</th><th>Meaning</th></tr>
+    <tr><td>0</td><td>Success</td></tr>
+    <tr><td>1</td><td>Internal error</td></tr>
+    <tr><td>2</td><td>Config not found</td></tr>
+    <tr><td>3</td><td>Config invalid</td></tr>
+    <tr><td>4</td><td>Token not set</td></tr>
+    <tr><td>5</td><td>GitLab auth failed</td></tr>
+    <tr><td>6</td><td>Resource not found</td></tr>
+    <tr><td>7</td><td>Rate limited</td></tr>
+    <tr><td>8</td><td>Network error</td></tr>
+    <tr><td>9</td><td>Database locked</td></tr>
+    <tr><td>10</td><td>Database error</td></tr>
+    <tr><td>11</td><td>Migration failed</td></tr>
+    <tr><td>12</td><td>I/O error</td></tr>
+    <tr><td>13</td><td>Transform error</td></tr>
+  </table>
+
+  <h3>Configuration</h3>
+  <pre>{
+  "gitlab": {
+    "baseUrl": "https://gitlab.com",
+    "tokenEnvVar": "GITLAB_TOKEN"
+  },
+  "projects": [
+    { "path": "group/project" }
+  ],
+  "sync": {
+    "backfillDays": 14,
+    "staleLockMinutes": 10,
+    "heartbeatIntervalSeconds": 30,
+    "cursorRewindSeconds": 2,
+    "primaryConcurrency": 4,
+    "dependentConcurrency": 2
+  },
+  "storage": {
+    "dbPath": "~/.local/share/lore/lore.db",
+    "compressRawPayloads": true
+  }
+}</pre>
+
+  <h3>Deprecated Commands (Hidden)</h3>
+  <table>
+    <tr><th>Old</th><th>New</th><th>Notes</th></tr>
+    <tr><td><code>lore list</code></td><td><code>lore issues</code> / <code>lore mrs</code></td><td>Shows deprecation warning</td></tr>
+    <tr><td><code>lore show</code></td><td><code>lore issues &lt;iid&gt;</code></td><td>Shows deprecation warning</td></tr>
+    <tr><td><code>lore auth-test</code></td><td><code>lore auth</code></td><td>Alias</td></tr>
+    <tr><td><code>lore sync-status</code></td><td><code>lore status</code></td><td>Alias</td></tr>
+  </table>
+</div>
+
+<!-- API ENDPOINTS USED -->
+<div id="api-used" class="panel">
+  <h2>GitLab API Endpoints Used by Gitlore</h2>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:16px">All requests use <code>PRIVATE-TOKEN</code> header authentication. Rate limited at 10 req/s with 0-50ms jitter.</p>
+
+  <div class="expandable" data-id="ep-issues">
+    <div class="expandable-header" onclick="toggleExpand('ep-issues')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/projects/:id/issues</code> &mdash; List project issues</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-issues-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore ingest issues</code></p>
+      <h4>Query Parameters Used</h4>
+      <table>
+        <tr><th>Param</th><th>Value</th><th>Purpose</th></tr>
+        <tr><td><code>scope</code></td><td><code>all</code></td><td>Include all issues, not just assigned to user</td></tr>
+        <tr><td><code>state</code></td><td><code>all</code></td><td>Open and closed</td></tr>
+        <tr><td><code>order_by</code></td><td><code>updated_at</code></td><td>Support cursor-based incremental sync</td></tr>
+        <tr><td><code>sort</code></td><td><code>asc</code></td><td>Oldest first for cursor correctness</td></tr>
+        <tr><td><code>per_page</code></td><td><code>100</code></td><td>Maximum page size</td></tr>
+        <tr><td><code>page</code></td><td><code>{n}</code></td><td>Offset pagination</td></tr>
+        <tr><td><code>updated_after</code></td><td>ISO 8601</td><td>Cursor: only fetch changes since last sync</td></tr>
+      </table>
+      <h4>Response Fields Consumed</h4>
+      <pre>id, iid, project_id, title, description, state,
+created_at, updated_at, closed_at,
+author { id, username, name },
+assignees[] { id, username, name },
+labels[], milestone { id, iid, title, state, due_date, web_url },
+web_url, due_date</pre>
+      <h4>Pagination</h4>
+      <p style="font-size:13px">Fallback chain: <code>Link</code> header (RFC 8288) &rarr; <code>x-next-page</code> header &rarr; full-page heuristic (100 items = more pages)</p>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="ep-mrs">
+    <div class="expandable-header" onclick="toggleExpand('ep-mrs')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/projects/:id/merge_requests</code> &mdash; List project MRs</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-mrs-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore ingest mrs</code></p>
+      <h4>Query Parameters Used</h4>
+      <table>
+        <tr><th>Param</th><th>Value</th><th>Purpose</th></tr>
+        <tr><td><code>scope</code></td><td><code>all</code></td><td>All MRs in project</td></tr>
+        <tr><td><code>state</code></td><td><code>all</code></td><td>opened, merged, closed, locked</td></tr>
+        <tr><td><code>order_by</code></td><td><code>updated_at</code></td><td>Cursor-based sync</td></tr>
+        <tr><td><code>sort</code></td><td><code>asc</code></td><td>Oldest first</td></tr>
+        <tr><td><code>per_page</code></td><td><code>100</code></td><td>Max page size</td></tr>
+        <tr><td><code>page</code></td><td><code>{n}</code></td><td>Offset pagination</td></tr>
+        <tr><td><code>updated_after</code></td><td>ISO 8601</td><td>Incremental cursor</td></tr>
+      </table>
+      <h4>Response Fields Consumed</h4>
+      <pre>id, iid, project_id, title, description, state,
+draft (preferred) OR work_in_progress (fallback),
+author, assignees[], reviewers[],
+labels[], source_branch, target_branch, sha,
+references { short, full },
+detailed_merge_status (preferred) OR merge_status (fallback),
+merge_user (preferred) OR merged_by (fallback),
+created_at, updated_at, merged_at, closed_at, web_url</pre>
+      <h4>Deprecated Field Fallbacks</h4>
+      <table>
+        <tr><th>Preferred</th><th>Fallback</th></tr>
+        <tr><td><code>draft</code></td><td><code>work_in_progress</code></td></tr>
+        <tr><td><code>detailed_merge_status</code></td><td><code>merge_status</code></td></tr>
+        <tr><td><code>merge_user</code></td><td><code>merged_by</code></td></tr>
+      </table>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="ep-issue-disc">
+    <div class="expandable-header" onclick="toggleExpand('ep-issue-disc')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/projects/:id/issues/:iid/discussions</code> &mdash; Issue discussions</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-issue-disc-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore ingest issues</code> (dependent phase)</p>
+      <h4>Sync Strategy</h4>
+      <p style="font-size:13px">Full-refresh: DELETE all existing discussions for the issue, INSERT all fetched discussions. Only triggered when <code>issue.updated_at &gt; issue.discussions_synced_for_updated_at</code>.</p>
+      <h4>Response Fields Consumed</h4>
+      <pre>id (string), individual_note (bool),
+notes[] {
+  id, type (DiscussionNote | DiffNote | null),
+  body, author { username, name },
+  created_at, updated_at,
+  system (bool), resolvable, resolved,
+  resolved_by, resolved_at,
+  position { old_path, new_path, old_line, new_line,
+    position_type, line_range, base_sha, start_sha, head_sha }
+}</pre>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="ep-mr-disc">
+    <div class="expandable-header" onclick="toggleExpand('ep-mr-disc')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/projects/:id/merge_requests/:iid/discussions</code> &mdash; MR discussions</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-mr-disc-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore ingest mrs</code> (dependent phase)</p>
+      <h4>Sync Strategy</h4>
+      <p style="font-size:13px">Same full-refresh strategy as issue discussions. Includes DiffNote support (code review comments on specific lines).</p>
+      <h4>Additional: DiffNote Position</h4>
+      <pre>position {
+  old_path, new_path,       // file paths
+  old_line, new_line,       // line numbers
+  position_type,            // "text" or "image"
+  line_range {              // multi-line comments
+    start { line_code, type, old_line, new_line },
+    end { line_code, type, old_line, new_line }
+  },
+  base_sha, start_sha, head_sha  // commit references
+}</pre>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="ep-user">
+    <div class="expandable-header" onclick="toggleExpand('ep-user')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/user</code> &mdash; Current authenticated user</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-user-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore auth</code>, <code>lore doctor</code></p>
+      <h4>Response Fields</h4>
+      <pre>id, username, name, email, avatar_url, web_url, created_at, state</pre>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="ep-project">
+    <div class="expandable-header" onclick="toggleExpand('ep-project')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/projects/:path_encoded</code> &mdash; Project details</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-project-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore ingest</code> (project resolution during init/sync)</p>
+      <h4>Response Fields</h4>
+      <pre>id, path_with_namespace, default_branch, web_url,
+created_at, updated_at, name, description, visibility, archived</pre>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="ep-version">
+    <div class="expandable-header" onclick="toggleExpand('ep-version')">
+      <span><span class="badge badge-get">GET</span> <code>/api/v4/version</code> &mdash; GitLab instance version</span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="ep-version-body">
+      <h4>Used By</h4>
+      <p style="font-size:13px"><code>lore doctor</code></p>
+      <h4>Response Fields</h4>
+      <pre>version, revision</pre>
+    </div>
+  </div>
+</div>
+
+<!-- FULL GITLAB API -->
+<div id="api-full" class="panel">
+  <h2>Full GitLab REST API v4 Reference</h2>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:16px">Complete endpoint inventory for the resources relevant to Gitlore. <span class="badge badge-used">USED</span> = consumed by Gitlore.</p>
+
+  <input type="text" class="search-box" id="api-search" placeholder="Filter endpoints... (e.g. 'issues', 'POST', 'discussions')" oninput="filterAPI()">
+
+  <div id="api-list">
+    <!-- ISSUES -->
+    <h3>Issues API</h3>
+    <table class="api-table">
+      <tr><th style="width:60px">Method</th><th>Endpoint</th><th>Description</th><th style="width:60px">Status</th></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/issues</code></td><td>List all issues (global)</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/groups/:id/issues</code></td><td>List group issues</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues</code></td><td>List project issues</td><td><span class="badge badge-used">USED</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid</code></td><td>Get single issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues</code></td><td>Create issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/issues/:iid</code></td><td>Update issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-delete">DEL</span></td><td><code>/projects/:id/issues/:iid</code></td><td>Delete issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/issues/:iid/reorder</code></td><td>Reorder issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/move</code></td><td>Move issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/clone</code></td><td>Clone issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/subscribe</code></td><td>Subscribe to issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/unsubscribe</code></td><td>Unsubscribe</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/todo</code></td><td>Create to-do</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/time_estimate</code></td><td>Set time estimate</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/add_spent_time</code></td><td>Add spent time</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/time_stats</code></td><td>Get time stats</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/related_merge_requests</code></td><td>Related MRs</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/closed_by</code></td><td>MRs that close issue</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/participants</code></td><td>List participants</td><td><span class="badge badge-unused">--</span></td></tr>
+    </table>
+
+    <!-- MERGE REQUESTS -->
+    <h3>Merge Requests API</h3>
+    <table class="api-table">
+      <tr><th style="width:60px">Method</th><th>Endpoint</th><th>Description</th><th style="width:60px">Status</th></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/merge_requests</code></td><td>List all MRs (global)</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/groups/:id/merge_requests</code></td><td>List group MRs</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests</code></td><td>List project MRs</td><td><span class="badge badge-used">USED</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid</code></td><td>Get single MR</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/merge_requests</code></td><td>Create MR</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/merge_requests/:iid</code></td><td>Update MR</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-delete">DEL</span></td><td><code>/projects/:id/merge_requests/:iid</code></td><td>Delete MR</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/merge_requests/:iid/merge</code></td><td>Merge an MR</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/merge_requests/:iid/cancel_merge</code></td><td>Cancel merge</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/merge_requests/:iid/rebase</code></td><td>Rebase MR</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/commits</code></td><td>List MR commits</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/changes</code></td><td>List MR diffs</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/pipelines</code></td><td>MR pipelines</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/participants</code></td><td>MR participants</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/approvals</code></td><td>MR approvals</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/merge_requests/:iid/approve</code></td><td>Approve MR</td><td><span class="badge badge-unused">--</span></td></tr>
+    </table>
+
+    <!-- DISCUSSIONS -->
+    <h3>Discussions API</h3>
+    <table class="api-table">
+      <tr><th style="width:60px">Method</th><th>Endpoint</th><th>Description</th><th style="width:60px">Status</th></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/discussions</code></td><td>List issue discussions</td><td><span class="badge badge-used">USED</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/discussions/:did</code></td><td>Get single discussion</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/discussions</code></td><td>Create issue thread</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/discussions/:did/notes</code></td><td>Add note to thread</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/issues/:iid/discussions/:did/notes/:nid</code></td><td>Modify note</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-delete">DEL</span></td><td><code>/projects/:id/issues/:iid/discussions/:did/notes/:nid</code></td><td>Delete note</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/discussions</code></td><td>List MR discussions</td><td><span class="badge badge-used">USED</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/discussions/:did</code></td><td>Get single MR discussion</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/merge_requests/:iid/discussions</code></td><td>Create MR thread</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/merge_requests/:iid/discussions/:did</code></td><td>Resolve/unresolve thread</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/merge_requests/:iid/discussions/:did/notes</code></td><td>Add note to MR thread</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-put">PUT</span></td><td><code>/projects/:id/merge_requests/:iid/discussions/:did/notes/:nid</code></td><td>Modify MR note</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-delete">DEL</span></td><td><code>/projects/:id/merge_requests/:iid/discussions/:did/notes/:nid</code></td><td>Delete MR note</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/snippets/:sid/discussions</code></td><td>List snippet discussions</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/groups/:id/epics/:eid/discussions</code></td><td>List epic discussions</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/repository/commits/:sha/discussions</code></td><td>List commit discussions</td><td><span class="badge badge-unused">--</span></td></tr>
+    </table>
+
+    <!-- NOTES -->
+    <h3>Notes API (Flat, non-threaded)</h3>
+    <table class="api-table">
+      <tr><th style="width:60px">Method</th><th>Endpoint</th><th>Description</th><th style="width:60px">Status</th></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/issues/:iid/notes</code></td><td>List issue notes</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/issues/:iid/notes</code></td><td>Create issue note</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/merge_requests/:iid/notes</code></td><td>List MR notes</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-post">POST</span></td><td><code>/projects/:id/merge_requests/:iid/notes</code></td><td>Create MR note</td><td><span class="badge badge-unused">--</span></td></tr>
+      <tr><td><span class="badge badge-get">GET</span></td><td><code>/projects/:id/snippets/:sid/notes</code></td><td>List snippet notes</td><td><span class="badge badge-unused">--</span></td></tr>
+    </table>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px">Gitlore uses the Discussions API (threaded) instead of the flat Notes API. Notes are extracted from discussion responses.</p>
+
+    <!-- OTHER USED -->
+    <h3>Other APIs Used</h3>
+    <table class="api-table">
+      <tr><th style="width:60px">Method</th><th>Endpoint</th><th>Description</th><th style="width:60px">Status</th></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/user</code></td><td>Current authenticated user</td><td><span class="badge badge-used">USED</span></td></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/projects/:path</code></td><td>Get project by path</td><td><span class="badge badge-used">USED</span></td></tr>
+      <tr class="used-row"><td><span class="badge badge-get">GET</span></td><td><code>/version</code></td><td>GitLab instance version</td><td><span class="badge badge-used">USED</span></td></tr>
+    </table>
+  </div>
+</div>
+
+<!-- MAPPING -->
+<div id="mapping" class="panel">
+  <h2>CLI Command &harr; API Endpoint Mapping</h2>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:16px">How each CLI command maps to GitLab API calls and local database operations.</p>
+
+  <div class="card">
+    <div class="card-title">lore ingest issues</div>
+    <div class="mapping-row">
+      <div><strong>Phase 1:</strong> Fetch primary resources</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /projects/:id/issues</code> (paginated, cursor-based)</div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Phase 2:</strong> Identify stale discussions</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>SQL: <code>WHERE updated_at > discussions_synced_for_updated_at</code></div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Phase 3:</strong> Sync discussions</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /projects/:id/issues/:iid/discussions</code> (parallel prefetch)</div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Storage:</strong> Write to DB</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>Tables: <code>issues</code>, <code>labels</code>, <code>issue_labels</code>, <code>discussions</code>, <code>notes</code>, <code>raw_payloads</code></div>
+    </div>
+  </div>
+
+  <div class="card">
+    <div class="card-title">lore ingest mrs</div>
+    <div class="mapping-row">
+      <div><strong>Phase 1:</strong> Fetch primary resources</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /projects/:id/merge_requests</code> (paginated, cursor-based)</div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Phase 2:</strong> Identify stale discussions</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>SQL: <code>WHERE updated_at > discussions_synced_for_updated_at</code></div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Phase 3:</strong> Sync discussions</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /projects/:id/merge_requests/:iid/discussions</code> (parallel prefetch)</div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Storage:</strong> Write to DB</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>Tables: <code>merge_requests</code>, <code>labels</code>, <code>mr_labels</code>, <code>mr_assignees</code>, <code>mr_reviewers</code>, <code>discussions</code>, <code>notes</code>, <code>raw_payloads</code></div>
+    </div>
+  </div>
+
+  <div class="card">
+    <div class="card-title">lore issues / lore mrs</div>
+    <div class="mapping-row">
+      <div><strong>List mode:</strong> Query local DB</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>SQL: <code>SELECT ... FROM issues/merge_requests</code> with filters (no API call)</div>
+    </div>
+    <div class="mapping-row">
+      <div><strong>Show mode:</strong> Query local DB by IID</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>SQL: <code>SELECT ... WHERE iid = ?</code> + join discussions/notes (no API call)</div>
+    </div>
+  </div>
+
+  <div class="card">
+    <div class="card-title">lore auth</div>
+    <div class="mapping-row">
+      <div>Verify token works</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /api/v4/user</code></div>
+    </div>
+  </div>
+
+  <div class="card">
+    <div class="card-title">lore doctor</div>
+    <div class="mapping-row">
+      <div>Check auth + GitLab version</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /api/v4/user</code> + <code>GET /api/v4/version</code></div>
+    </div>
+    <div class="mapping-row">
+      <div>Check each configured project</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div><code>GET /api/v4/projects/:path</code></div>
+    </div>
+  </div>
+
+  <div class="card">
+    <div class="card-title">lore count / lore status / lore init / lore migrate</div>
+    <div class="mapping-row">
+      <div>Local-only operations</div>
+      <div class="mapping-connector">&rarr;</div>
+      <div>No API calls. Database queries only.</div>
+    </div>
+  </div>
+
+  <h3 style="margin-top:24px">API Capabilities NOT Used by Gitlore</h3>
+  <div class="grid-2">
+    <div class="card">
+      <div class="card-title" style="color:var(--red)">Write Operations</div>
+      <ul style="font-size:13px;padding-left:16px;color:var(--text-muted)">
+        <li>Create/update/delete issues</li>
+        <li>Create/update/delete MRs</li>
+        <li>Merge MRs</li>
+        <li>Create/reply to discussions</li>
+        <li>Resolve/unresolve threads</li>
+        <li>Approve MRs</li>
+      </ul>
+    </div>
+    <div class="card">
+      <div class="card-title" style="color:var(--yellow)">Read Operations</div>
+      <ul style="font-size:13px;padding-left:16px;color:var(--text-muted)">
+        <li>Single issue/MR fetch (uses list with filters instead)</li>
+        <li>MR commits, diffs, pipelines</li>
+        <li>Issue/MR participants</li>
+        <li>Time tracking stats</li>
+        <li>Related MRs / closed-by</li>
+        <li>Labels API (extracted from issue/MR responses)</li>
+        <li>Milestones API (extracted from issue responses)</li>
+        <li>Flat Notes API (uses threaded Discussions API)</li>
+        <li>Snippets, Epics, Commits discussions</li>
+        <li>Webhooks, CI/CD, Pipelines, Deployments</li>
+      </ul>
+    </div>
+  </div>
+</div>
+
+<!-- DATA MODEL -->
+<div id="data" class="panel">
+  <h2>Database Schema</h2>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:16px">SQLite with WAL mode. 12 tables across 6 migrations.</p>
+
+  <h3>Entity Relationship</h3>
+  <pre style="font-size:11px;line-height:1.4">
+  projects ──────────────────────────────────────────────────────────
+    │                                                                │
+    ├──&lt; issues ──&lt; issue_labels &gt;── labels                          │
+    │     │                                                          │
+    │     └──&lt; discussions ──&lt; notes                                 │
+    │                                                                │
+    ├──&lt; merge_requests ──&lt; mr_labels &gt;── labels                    │
+    │     │    │    │                                                 │
+    │     │    │    └──&lt; mr_reviewers                                 │
+    │     │    └──&lt; mr_assignees                                     │
+    │     │                                                          │
+    │     └──&lt; discussions ──&lt; notes                                 │
+    │                                                                │
+    ├──&lt; raw_payloads                                                │
+    ├──&lt; sync_cursors                                                │
+    └── sync_runs, app_locks, schema_version                         │
+  ───────────────────────────────────────────────────────────────────</pre>
+
+  <h3>Table Details</h3>
+
+  <div class="expandable" data-id="tbl-projects">
+    <div class="expandable-header" onclick="toggleExpand('tbl-projects')">
+      <span><strong>projects</strong> <span style="color:var(--text-muted);font-size:12px">&mdash; Configured GitLab projects</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="tbl-projects-body">
+      <table>
+        <tr><th>Column</th><th>Type</th><th>Notes</th></tr>
+        <tr><td>id</td><td>INTEGER PK</td><td>Auto-increment</td></tr>
+        <tr><td>gitlab_project_id</td><td>INTEGER UNIQUE</td><td>GitLab's project ID</td></tr>
+        <tr><td>path_with_namespace</td><td>TEXT</td><td>e.g. "group/project"</td></tr>
+        <tr><td>default_branch</td><td>TEXT</td><td>e.g. "main"</td></tr>
+        <tr><td>web_url</td><td>TEXT</td><td>Full URL</td></tr>
+        <tr><td>created_at, updated_at</td><td>INTEGER</td><td>ms epoch</td></tr>
+        <tr><td>raw_payload_id</td><td>INTEGER FK</td><td>Link to raw_payloads</td></tr>
+      </table>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="tbl-issues">
+    <div class="expandable-header" onclick="toggleExpand('tbl-issues')">
+      <span><strong>issues</strong> <span style="color:var(--text-muted);font-size:12px">&mdash; Synced GitLab issues</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="tbl-issues-body">
+      <table>
+        <tr><th>Column</th><th>Type</th><th>Notes</th></tr>
+        <tr><td>id</td><td>INTEGER PK</td><td>Auto-increment</td></tr>
+        <tr><td>gitlab_id</td><td>INTEGER UNIQUE</td><td>GitLab's issue ID</td></tr>
+        <tr><td>project_id</td><td>INTEGER FK</td><td>projects.id</td></tr>
+        <tr><td>iid</td><td>INTEGER</td><td>Issue # within project (UNIQUE with project_id)</td></tr>
+        <tr><td>title, description</td><td>TEXT</td><td></td></tr>
+        <tr><td>state</td><td>TEXT</td><td>'opened' | 'closed'</td></tr>
+        <tr><td>author_username</td><td>TEXT</td><td>Denormalized</td></tr>
+        <tr><td>created_at, updated_at, closed_at</td><td>INTEGER</td><td>ms epoch</td></tr>
+        <tr><td>last_seen_at</td><td>INTEGER</td><td>Orphan detection</td></tr>
+        <tr><td>discussions_synced_for_updated_at</td><td>INTEGER</td><td>Discussion sync watermark</td></tr>
+        <tr><td>web_url</td><td>TEXT</td><td></td></tr>
+        <tr><td>raw_payload_id</td><td>INTEGER FK</td><td></td></tr>
+      </table>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="tbl-mrs">
+    <div class="expandable-header" onclick="toggleExpand('tbl-mrs')">
+      <span><strong>merge_requests</strong> <span style="color:var(--text-muted);font-size:12px">&mdash; Synced GitLab MRs</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="tbl-mrs-body">
+      <table>
+        <tr><th>Column</th><th>Type</th><th>Notes</th></tr>
+        <tr><td>id</td><td>INTEGER PK</td><td></td></tr>
+        <tr><td>gitlab_id</td><td>INTEGER UNIQUE</td><td></td></tr>
+        <tr><td>project_id</td><td>INTEGER FK</td><td></td></tr>
+        <tr><td>iid</td><td>INTEGER</td><td></td></tr>
+        <tr><td>title, description, state</td><td>TEXT</td><td>'opened' | 'merged' | 'closed' | 'locked'</td></tr>
+        <tr><td>draft</td><td>INTEGER</td><td>0/1</td></tr>
+        <tr><td>author_username</td><td>TEXT</td><td></td></tr>
+        <tr><td>source_branch, target_branch</td><td>TEXT</td><td></td></tr>
+        <tr><td>head_sha</td><td>TEXT</td><td></td></tr>
+        <tr><td>references_short, references_full</td><td>TEXT</td><td>e.g. "!42", "group/proj!42"</td></tr>
+        <tr><td>detailed_merge_status</td><td>TEXT</td><td></td></tr>
+        <tr><td>merge_user_username</td><td>TEXT</td><td></td></tr>
+        <tr><td>created_at, updated_at, merged_at, closed_at</td><td>INTEGER</td><td>ms epoch</td></tr>
+        <tr><td>last_seen_at</td><td>INTEGER</td><td></td></tr>
+        <tr><td>discussions_synced_for_updated_at</td><td>INTEGER</td><td>Watermark</td></tr>
+        <tr><td>discussions_sync_last_attempt_at</td><td>INTEGER</td><td>Health telemetry</td></tr>
+        <tr><td>discussions_sync_attempts</td><td>INTEGER</td><td>Retry count</td></tr>
+        <tr><td>discussions_sync_last_error</td><td>TEXT</td><td></td></tr>
+        <tr><td>web_url, raw_payload_id</td><td>TEXT/INT</td><td></td></tr>
+      </table>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="tbl-discussions">
+    <div class="expandable-header" onclick="toggleExpand('tbl-discussions')">
+      <span><strong>discussions + notes</strong> <span style="color:var(--text-muted);font-size:12px">&mdash; Threaded comments</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="tbl-discussions-body">
+      <h4>discussions</h4>
+      <table>
+        <tr><th>Column</th><th>Type</th><th>Notes</th></tr>
+        <tr><td>id</td><td>INTEGER PK</td><td></td></tr>
+        <tr><td>gitlab_discussion_id</td><td>TEXT</td><td>SHA-like string ID</td></tr>
+        <tr><td>project_id</td><td>INTEGER FK</td><td></td></tr>
+        <tr><td>issue_id</td><td>INTEGER FK</td><td>XOR with merge_request_id</td></tr>
+        <tr><td>merge_request_id</td><td>INTEGER FK</td><td>XOR with issue_id</td></tr>
+        <tr><td>noteable_type</td><td>TEXT</td><td>'Issue' | 'MergeRequest'</td></tr>
+        <tr><td>individual_note</td><td>INTEGER</td><td>0 = threaded, 1 = standalone</td></tr>
+        <tr><td>resolvable, resolved</td><td>INTEGER</td><td>0/1</td></tr>
+        <tr><td>first_note_at, last_note_at</td><td>INTEGER</td><td>Computed from notes</td></tr>
+      </table>
+      <h4>notes</h4>
+      <table>
+        <tr><th>Column</th><th>Type</th><th>Notes</th></tr>
+        <tr><td>id</td><td>INTEGER PK</td><td></td></tr>
+        <tr><td>gitlab_id</td><td>INTEGER UNIQUE</td><td></td></tr>
+        <tr><td>discussion_id</td><td>INTEGER FK</td><td></td></tr>
+        <tr><td>note_type</td><td>TEXT</td><td>'DiscussionNote' | 'DiffNote' | null</td></tr>
+        <tr><td>is_system</td><td>INTEGER</td><td>Auto-generated (label changes, etc.)</td></tr>
+        <tr><td>author_username, body</td><td>TEXT</td><td></td></tr>
+        <tr><td>position</td><td>INTEGER</td><td>0-indexed order within discussion</td></tr>
+        <tr><td>resolvable, resolved</td><td>INTEGER</td><td>0/1</td></tr>
+        <tr><td>resolved_by, resolved_at</td><td>TEXT/INT</td><td></td></tr>
+        <tr><td>position_old/new_path</td><td>TEXT</td><td>DiffNote: file path</td></tr>
+        <tr><td>position_old/new_line</td><td>INTEGER</td><td>DiffNote: line number</td></tr>
+        <tr><td>position_type</td><td>TEXT</td><td>DiffNote: "text" or "image"</td></tr>
+        <tr><td>position_line_range_start/end</td><td>TEXT</td><td>JSON for multi-line</td></tr>
+        <tr><td>position_base/start/head_sha</td><td>TEXT</td><td>Commit references</td></tr>
+      </table>
+    </div>
+  </div>
+
+  <div class="expandable" data-id="tbl-infra">
+    <div class="expandable-header" onclick="toggleExpand('tbl-infra')">
+      <span><strong>Infrastructure tables</strong> <span style="color:var(--text-muted);font-size:12px">&mdash; sync_cursors, sync_runs, app_locks, raw_payloads</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="tbl-infra-body">
+      <h4>sync_cursors</h4>
+      <p style="font-size:13px">PK: <code>(project_id, resource_type)</code>. Tracks <code>updated_at_cursor</code> (ms epoch) and <code>tie_breaker_id</code> for incremental sync.</p>
+
+      <h4>sync_runs</h4>
+      <p style="font-size:13px">Audit trail: <code>started_at</code>, <code>finished_at</code>, <code>status</code> ('running'|'succeeded'|'failed'), <code>command</code>, <code>error</code>, <code>metrics_json</code>.</p>
+
+      <h4>app_locks</h4>
+      <p style="font-size:13px">Single-flight mutex: <code>name</code> PK ('sync'), <code>owner</code> (UUIDv4), <code>heartbeat_at</code>. Prevents concurrent sync.</p>
+
+      <h4>raw_payloads</h4>
+      <p style="font-size:13px">Archived API responses. SHA-256 dedup. Optional gzip. <code>UNIQUE(project_id, resource_type, gitlab_id, payload_hash)</code>.</p>
+
+      <h4>SQLite Pragmas</h4>
+      <pre>journal_mode = WAL          -- Write-ahead logging
+synchronous = NORMAL        -- Safe for WAL
+foreign_keys = ON           -- Referential integrity
+busy_timeout = 5000         -- 5s lock wait
+temp_store = MEMORY         -- Speed optimization</pre>
+    </div>
+  </div>
+</div>
+
+<!-- INGESTION PIPELINE -->
+<div id="pipeline" class="panel">
+  <h2>Ingestion Pipeline</h2>
+
+  <h3>Three-Phase Architecture</h3>
+  <div class="card">
+    <div style="display:flex;gap:16px;flex-wrap:wrap">
+      <div class="flow-box highlight" style="flex:1;min-width:200px">
+        <strong>Phase 1: Primary Fetch</strong><br>
+        <small style="color:var(--text-muted)">Paginated API fetch with cursor-based sync.<br>Stores raw payloads + normalized rows.</small>
+      </div>
+      <div class="flow-box" style="flex:1;min-width:200px">
+        <strong>Phase 2: Identify Stale</strong><br>
+        <small style="color:var(--text-muted)">SQL query: which issues/MRs need<br>their discussions refreshed?</small>
+      </div>
+      <div class="flow-box highlight" style="flex:1;min-width:200px">
+        <strong>Phase 3: Discussion Sync</strong><br>
+        <small style="color:var(--text-muted)">Parallel prefetch + serial write.<br>Full-refresh per parent entity.</small>
+      </div>
+    </div>
+  </div>
+
+  <h3>Cursor-Based Incremental Sync</h3>
+  <div class="card">
+    <pre style="margin:0">Cursor State: (updated_at_cursor: i64, tie_breaker_id: i64)
+
+First sync:
+  updated_after = (now - backfillDays)
+
+Subsequent syncs:
+  updated_after = cursor.updated_at - cursorRewindSeconds
+
+  For each fetched resource:
+    if (gitlab_id, updated_at) <= cursor:
+      SKIP (already processed in overlap zone)
+    else:
+      UPSERT into database
+
+  After each page boundary:
+    UPDATE sync_cursors  (crash recovery safe)</pre>
+  </div>
+
+  <h3>Discussion Sync Strategy</h3>
+  <div class="card">
+    <pre style="margin:0">For each issue/MR where updated_at > discussions_synced_for_updated_at:
+
+  1. PREFETCH (parallel, configurable concurrency):
+     GET /projects/:id/issues/:iid/discussions  (all pages)
+
+  2. WRITE (serial, inside transaction):
+     DELETE FROM discussions WHERE issue_id = ?
+     DELETE FROM notes WHERE discussion_id IN (...)
+     INSERT discussions + notes (fresh data)
+     UPDATE issues SET discussions_synced_for_updated_at = updated_at</pre>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px">Full-refresh avoids complexity of detecting deleted/edited notes. Trade-off: more API calls for heavily-discussed items.</p>
+  </div>
+
+  <h3>Rate Limiting</h3>
+  <div class="card">
+    <pre style="margin:0">RateLimiter {
+  min_interval: 100ms  (= 1s / 10 req/s)
+  jitter: 0-50ms random
+
+  acquire():
+    elapsed = now - last_request
+    if elapsed < min_interval:
+      sleep(min_interval - elapsed + random_jitter)
+    last_request = now
+}</pre>
+  </div>
+
+  <h3>Pagination</h3>
+  <div class="card">
+    <p style="font-size:13px">Async stream-based. Fallback chain for next-page detection:</p>
+    <ol style="font-size:13px;padding-left:20px;margin-top:8px">
+      <li><code>Link</code> header (RFC 8288) &mdash; parse <code>rel="next"</code></li>
+      <li><code>x-next-page</code> header &mdash; direct page number</li>
+      <li>Full-page heuristic &mdash; if response has 100 items, assume more pages</li>
+    </ol>
+  </div>
+
+  <h3>Raw Payload Storage</h3>
+  <div class="card">
+    <div class="flow-diagram">
+      <div class="flow-box">API Response<br><small>JSON bytes</small></div>
+      <div class="flow-arrow">&rarr;</div>
+      <div class="flow-box">SHA-256 Hash<br><small>Dedup check</small></div>
+      <div class="flow-arrow">&rarr;</div>
+      <div class="flow-box">Gzip Compress<br><small>(if enabled)</small></div>
+      <div class="flow-arrow">&rarr;</div>
+      <div class="flow-box highlight">raw_payloads<br><small>BLOB storage</small></div>
+    </div>
+    <p style="font-size:12px;color:var(--text-muted)">UNIQUE constraint on <code>(project_id, resource_type, gitlab_id, payload_hash)</code> prevents storing identical payloads.</p>
+  </div>
+
+  <h3>Concurrency Model</h3>
+  <div class="grid-2">
+    <div class="card">
+      <div class="card-title">Primary Resource Fetch</div>
+      <p style="font-size:13px;color:var(--text-muted)">Single-threaded async stream. Rate-limited. Each page written in a transaction. Cursor updated at page boundaries.</p>
+    </div>
+    <div class="card">
+      <div class="card-title">Discussion Sync</div>
+      <p style="font-size:13px;color:var(--text-muted)">Parallel prefetch (configurable, default 2 concurrent). Serial write phase to avoid DB contention. Each parent entity is one transaction.</p>
+    </div>
+  </div>
+
+  <h3>Single-Flight Lock</h3>
+  <div class="card">
+    <pre style="margin:0">AppLock (database-enforced mutex):
+  name: 'sync' (PK)
+  owner: UUIDv4 (unique per process)
+  heartbeat_at: updated every 30s
+
+  Acquire:
+    INSERT OR fail if row exists
+    Check stale: if heartbeat > staleLockMinutes, force-acquire
+
+  Release:
+    DELETE WHERE owner = my_uuid</pre>
+  </div>
+
+  <h3>Progress Events</h3>
+  <table>
+    <tr><th>Event</th><th>Description</th></tr>
+    <tr><td><code>IssuesFetchStarted</code></td><td>Beginning primary issue fetch</td></tr>
+    <tr><td><code>IssueFetched</code></td><td>Each issue processed (for progress bars)</td></tr>
+    <tr><td><code>IssuesFetchComplete</code></td><td>All pages consumed</td></tr>
+    <tr><td><code>DiscussionSyncStarted</code></td><td>Beginning discussion phase</td></tr>
+    <tr><td><code>DiscussionSynced</code></td><td>Each parent's discussions written</td></tr>
+    <tr><td><code>DiscussionSyncComplete</code></td><td>All discussions updated</td></tr>
+    <tr><td colspan="2" style="color:var(--text-muted);font-size:12px">Same events exist for MRs (MrsFetchStarted, etc.)</td></tr>
+  </table>
+</div>
+
+<!-- FIELD COVERAGE -->
+<div id="fields" class="panel">
+  <h2>Field-Level Coverage: API Response vs Gitlore Storage</h2>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:12px">Every field in every GitLab API response, mapped to what Gitlore does with it. Serde silently drops fields not in the Rust structs.</p>
+
+  <div class="legend">
+    <div class="legend-item"><div class="legend-dot" style="background:var(--green)"></div> Stored in DB</div>
+    <div class="legend-item"><div class="legend-dot" style="background:var(--yellow)"></div> Used transiently (logic only)</div>
+    <div class="legend-item"><div class="legend-dot" style="background:var(--purple)"></div> Deserialized but ignored</div>
+    <div class="legend-item"><div class="legend-dot" style="background:var(--red)"></div> Never deserialized (silently dropped)</div>
+  </div>
+
+  <input type="text" class="search-box" id="field-search" placeholder="Filter fields... (e.g. 'upvotes', 'stored', 'dropped')" oninput="filterFields()">
+
+  <!-- ISSUES RESPONSE -->
+  <div class="expandable" data-id="fld-issue">
+    <div class="expandable-header open" onclick="toggleExpand('fld-issue')">
+      <span><strong>GET /projects/:id/issues</strong> &mdash; Issue response object <span style="color:var(--text-muted)">(15 stored / 30+ total)</span></span>
+      <span class="arrow" style="transform:rotate(90deg)">&#9654;</span>
+    </div>
+    <div class="expandable-body open" id="fld-issue-body">
+      <table class="field-table">
+        <tr><th style="width:200px">API Field</th><th style="width:90px">Status</th><th style="width:180px">DB Column</th><th>Notes</th></tr>
+        <tr class="f-stored"><td><code>id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.gitlab_id</code></td><td>GitLab's global issue ID</td></tr>
+        <tr class="f-stored"><td><code>iid</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.iid</code></td><td>Project-scoped #</td></tr>
+        <tr class="f-stored"><td><code>project_id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.project_id</code></td><td>FK to projects</td></tr>
+        <tr class="f-stored"><td><code>title</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.title</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>description</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.description</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>state</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.state</code></td><td>"opened" | "closed"</td></tr>
+        <tr class="f-stored"><td><code>created_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.created_at</code></td><td>Converted to ms epoch</td></tr>
+        <tr class="f-stored"><td><code>updated_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.updated_at</code></td><td>Converted to ms epoch; drives cursor</td></tr>
+        <tr class="f-stored"><td><code>closed_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.closed_at</code></td><td>Nullable, ms epoch</td></tr>
+        <tr class="f-stored"><td><code>due_date</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.due_date</code></td><td>YYYY-MM-DD string</td></tr>
+        <tr class="f-stored"><td><code>web_url</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.web_url</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>author.username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issues.author_username</code></td><td>Only username kept</td></tr>
+        <tr class="f-stored"><td><code>assignees[].username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issue_assignees</code></td><td>Junction table, username only</td></tr>
+        <tr class="f-stored"><td><code>labels[]</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>issue_labels</code> &rarr; <code>labels</code></td><td>Junction table</td></tr>
+        <tr class="f-stored"><td><code>milestone.*</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>milestones</code> table + denorm</td><td>id, iid, title, state, due_date, web_url</td></tr>
+        <tr class="f-ignored"><td><code>author.id</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>author.name</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>author.state</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>author.avatar_url</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>author.web_url</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-dropped"><td><code>type</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>"ISSUE" &mdash; not deserialized</td></tr>
+        <tr class="f-dropped"><td><code>upvotes</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Engagement metric, not captured</td></tr>
+        <tr class="f-dropped"><td><code>downvotes</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Engagement metric, not captured</td></tr>
+        <tr class="f-dropped"><td><code>merge_requests_count</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Count of linked MRs</td></tr>
+        <tr class="f-dropped"><td><code>user_notes_count</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Comment count</td></tr>
+        <tr class="f-dropped"><td><code>subscribed</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>User-specific subscription state</td></tr>
+        <tr class="f-dropped"><td><code>confidential</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Visibility flag</td></tr>
+        <tr class="f-dropped"><td><code>discussion_locked</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Lock state</td></tr>
+        <tr class="f-dropped"><td><code>weight</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Premium: issue weight</td></tr>
+        <tr class="f-dropped"><td><code>time_stats.*</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>time_estimate, total_time_spent, human_*</td></tr>
+        <tr class="f-dropped"><td><code>task_completion_status</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>{count, completed_count}</td></tr>
+        <tr class="f-dropped"><td><code>references.*</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>{short, relative, full}</td></tr>
+        <tr class="f-dropped"><td><code>closed_by</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>User who closed it</td></tr>
+        <tr class="f-dropped"><td><code>assignee</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Deprecated singular; uses assignees[]</td></tr>
+        <tr class="f-dropped"><td><code>has_tasks</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Task list presence</td></tr>
+        <tr class="f-dropped"><td><code>issue_type</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>"issue" | "incident" | "test_case"</td></tr>
+        <tr class="f-dropped"><td><code>severity</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Incident severity</td></tr>
+        <tr class="f-dropped"><td><code>imported</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Import flag</td></tr>
+        <tr class="f-dropped"><td><code>imported_from</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Import source</td></tr>
+        <tr class="f-dropped"><td><code>moved_to_id</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Target if moved</td></tr>
+        <tr class="f-dropped"><td><code>_links</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>HATEOAS links</td></tr>
+        <tr class="f-dropped"><td><code>epic</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Premium: parent epic</td></tr>
+        <tr class="f-dropped"><td><code>iteration</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Premium: iteration</td></tr>
+        <tr class="f-dropped"><td><code>health_status</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Ultimate: health</td></tr>
+      </table>
+    </div>
+  </div>
+
+  <!-- MR RESPONSE -->
+  <div class="expandable" data-id="fld-mr">
+    <div class="expandable-header" onclick="toggleExpand('fld-mr')">
+      <span><strong>GET /projects/:id/merge_requests</strong> &mdash; MR response object <span style="color:var(--text-muted)">(20 stored / 45+ total)</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="fld-mr-body">
+      <table class="field-table">
+        <tr><th style="width:220px">API Field</th><th style="width:90px">Status</th><th style="width:200px">DB Column</th><th>Notes</th></tr>
+        <tr class="f-stored"><td><code>id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.gitlab_id</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>iid</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.iid</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>project_id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.project_id</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>title</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.title</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>description</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.description</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>state</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.state</code></td><td>opened|merged|closed|locked</td></tr>
+        <tr class="f-stored"><td><code>draft</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.draft</code></td><td>Preferred field</td></tr>
+        <tr class="f-stored"><td><code>source_branch</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.source_branch</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>target_branch</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.target_branch</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>sha</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.head_sha</code></td><td>Head commit SHA</td></tr>
+        <tr class="f-stored"><td><code>references.short</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.references_short</code></td><td>e.g. "!42"</td></tr>
+        <tr class="f-stored"><td><code>references.full</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.references_full</code></td><td>e.g. "group/proj!42"</td></tr>
+        <tr class="f-stored"><td><code>detailed_merge_status</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.detailed_merge_status</code></td><td>Preferred field</td></tr>
+        <tr class="f-stored"><td><code>author.username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.author_username</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>merge_user.username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.merge_user_username</code></td><td>Preferred over merged_by</td></tr>
+        <tr class="f-stored"><td><code>assignees[].username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>mr_assignees</code></td><td>Junction table</td></tr>
+        <tr class="f-stored"><td><code>reviewers[].username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>mr_reviewers</code></td><td>Junction table</td></tr>
+        <tr class="f-stored"><td><code>labels[]</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>mr_labels</code> &rarr; <code>labels</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>created_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.created_at</code></td><td>ms epoch</td></tr>
+        <tr class="f-stored"><td><code>updated_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.updated_at</code></td><td>ms epoch; drives cursor</td></tr>
+        <tr class="f-stored"><td><code>merged_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.merged_at</code></td><td>ms epoch, nullable</td></tr>
+        <tr class="f-stored"><td><code>closed_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.closed_at</code></td><td>ms epoch, nullable</td></tr>
+        <tr class="f-stored"><td><code>web_url</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>merge_requests.web_url</code></td><td></td></tr>
+        <tr class="f-transient"><td><code>work_in_progress</code></td><td><span class="badge" style="background:#2a2a1a;color:var(--yellow)">TRANSIENT</span></td><td>&mdash;</td><td>Deprecated fallback for <code>draft</code></td></tr>
+        <tr class="f-transient"><td><code>merge_status</code></td><td><span class="badge" style="background:#2a2a1a;color:var(--yellow)">TRANSIENT</span></td><td>&mdash;</td><td>Deprecated fallback for <code>detailed_merge_status</code></td></tr>
+        <tr class="f-transient"><td><code>merged_by</code></td><td><span class="badge" style="background:#2a2a1a;color:var(--yellow)">TRANSIENT</span></td><td>&mdash;</td><td>Deprecated fallback for <code>merge_user</code></td></tr>
+        <tr class="f-dropped"><td><code>upvotes</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Engagement metric</td></tr>
+        <tr class="f-dropped"><td><code>downvotes</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Engagement metric</td></tr>
+        <tr class="f-dropped"><td><code>user_notes_count</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Comment count</td></tr>
+        <tr class="f-dropped"><td><code>source_project_id</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Fork source</td></tr>
+        <tr class="f-dropped"><td><code>target_project_id</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Fork target</td></tr>
+        <tr class="f-dropped"><td><code>milestone</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Full milestone object (stored for issues, not MRs)</td></tr>
+        <tr class="f-dropped"><td><code>merge_when_pipeline_succeeds</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Auto-merge flag</td></tr>
+        <tr class="f-dropped"><td><code>merge_commit_sha</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Merge commit reference</td></tr>
+        <tr class="f-dropped"><td><code>squash_commit_sha</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Squash commit reference</td></tr>
+        <tr class="f-dropped"><td><code>discussion_locked</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Lock state</td></tr>
+        <tr class="f-dropped"><td><code>should_remove_source_branch</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Cleanup preference</td></tr>
+        <tr class="f-dropped"><td><code>force_remove_source_branch</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Forced cleanup</td></tr>
+        <tr class="f-dropped"><td><code>squash</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Squash flag</td></tr>
+        <tr class="f-dropped"><td><code>squash_on_merge</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Squash on merge flag</td></tr>
+        <tr class="f-dropped"><td><code>has_conflicts</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Merge conflict state</td></tr>
+        <tr class="f-dropped"><td><code>blocking_discussions_resolved</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>All blocking threads resolved?</td></tr>
+        <tr class="f-dropped"><td><code>time_stats.*</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Time tracking data</td></tr>
+        <tr class="f-dropped"><td><code>task_completion_status</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>{count, completed_count}</td></tr>
+        <tr class="f-dropped"><td><code>closed_by</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>User who closed</td></tr>
+        <tr class="f-dropped"><td><code>prepared_at</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Preparation timestamp</td></tr>
+        <tr class="f-dropped"><td><code>merge_after</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Scheduled merge time</td></tr>
+        <tr class="f-dropped"><td><code>imported</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Import flag</td></tr>
+        <tr class="f-dropped"><td><code>approvals_before_merge</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Deprecated approval count</td></tr>
+        <tr class="f-dropped"><td><code>references.relative</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Only short + full stored</td></tr>
+      </table>
+    </div>
+  </div>
+
+  <!-- DISCUSSION RESPONSE -->
+  <div class="expandable" data-id="fld-disc">
+    <div class="expandable-header" onclick="toggleExpand('fld-disc')">
+      <span><strong>GET /projects/:id/.../discussions</strong> &mdash; Discussion + Note response <span style="color:var(--text-muted)">(18 stored / 28+ total)</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="fld-disc-body">
+      <h4>Discussion object</h4>
+      <table class="field-table">
+        <tr><th style="width:200px">API Field</th><th style="width:90px">Status</th><th style="width:180px">DB Column</th><th>Notes</th></tr>
+        <tr class="f-stored"><td><code>id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>discussions.gitlab_discussion_id</code></td><td>SHA-like string</td></tr>
+        <tr class="f-stored"><td><code>individual_note</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>discussions.individual_note</code></td><td>0/1</td></tr>
+        <tr class="f-stored"><td><code>notes[]</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes</code> table</td><td>Each note stored individually</td></tr>
+      </table>
+      <p style="font-size:12px;color:var(--text-muted);margin:8px 0">Computed from notes: <code>first_note_at</code>, <code>last_note_at</code>, <code>resolvable</code> (any note resolvable), <code>resolved</code> (all resolvable notes resolved)</p>
+
+      <h4>Note object (within discussion)</h4>
+      <table class="field-table">
+        <tr><th style="width:200px">API Field</th><th style="width:90px">Status</th><th style="width:180px">DB Column</th><th>Notes</th></tr>
+        <tr class="f-stored"><td><code>id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.gitlab_id</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>type</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.note_type</code></td><td>DiscussionNote|DiffNote|null</td></tr>
+        <tr class="f-stored"><td><code>body</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.body</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>system</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.is_system</code></td><td>Auto-generated notes</td></tr>
+        <tr class="f-stored"><td><code>author.username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.author_username</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>created_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.created_at</code></td><td>ms epoch</td></tr>
+        <tr class="f-stored"><td><code>updated_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.updated_at</code></td><td>ms epoch</td></tr>
+        <tr class="f-stored"><td><code>resolvable</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.resolvable</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>resolved</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.resolved</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>resolved_by.username</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.resolved_by</code></td><td>Username only</td></tr>
+        <tr class="f-stored"><td><code>resolved_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.resolved_at</code></td><td>ms epoch</td></tr>
+        <tr class="f-stored"><td><code>position.old_path</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_old_path</code></td><td>DiffNote only</td></tr>
+        <tr class="f-stored"><td><code>position.new_path</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_new_path</code></td><td>DiffNote only</td></tr>
+        <tr class="f-stored"><td><code>position.old_line</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_old_line</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>position.new_line</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_new_line</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>position.position_type</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_type</code></td><td>"text"|"image"|"file"</td></tr>
+        <tr class="f-stored"><td><code>position.line_range</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_line_range_*</code></td><td>Start/end line numbers extracted</td></tr>
+        <tr class="f-stored"><td><code>position.base_sha</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_base_sha</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>position.start_sha</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_start_sha</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>position.head_sha</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>notes.position_head_sha</code></td><td></td></tr>
+        <tr class="f-dropped"><td><code>attachment</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>File attachment metadata</td></tr>
+        <tr class="f-dropped"><td><code>noteable_id</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Parent ID (redundant, known from URL)</td></tr>
+        <tr class="f-dropped"><td><code>noteable_type</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Parent type (redundant)</td></tr>
+        <tr class="f-dropped"><td><code>noteable_iid</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Parent IID (redundant)</td></tr>
+        <tr class="f-dropped"><td><code>project_id</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>From note context (redundant)</td></tr>
+        <tr class="f-dropped"><td><code>commit_id</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>DiffNote: specific commit</td></tr>
+        <tr class="f-dropped"><td><code>position.width</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Image position only</td></tr>
+        <tr class="f-dropped"><td><code>position.height</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Image position only</td></tr>
+        <tr class="f-dropped"><td><code>position.x</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Image position only</td></tr>
+        <tr class="f-dropped"><td><code>position.y</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Image position only</td></tr>
+        <tr class="f-dropped"><td><code>suggestions[]</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Code suggestion diffs (from_line, to_line, content)</td></tr>
+        <tr class="f-dropped"><td><code>line_range.*.line_code</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Internal line identifier</td></tr>
+        <tr class="f-dropped"><td><code>line_range.*.type</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>"old"|"new" side indicator</td></tr>
+      </table>
+    </div>
+  </div>
+
+  <!-- PROJECT RESPONSE -->
+  <div class="expandable" data-id="fld-proj">
+    <div class="expandable-header" onclick="toggleExpand('fld-proj')">
+      <span><strong>GET /projects/:path</strong> &mdash; Project response <span style="color:var(--text-muted)">(6 stored / 40+ total)</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="fld-proj-body">
+      <table class="field-table">
+        <tr><th style="width:220px">API Field</th><th style="width:90px">Status</th><th style="width:200px">DB Column</th><th>Notes</th></tr>
+        <tr class="f-stored"><td><code>id</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>projects.gitlab_project_id</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>path_with_namespace</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>projects.path_with_namespace</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>default_branch</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>projects.default_branch</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>web_url</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>projects.web_url</code></td><td></td></tr>
+        <tr class="f-stored"><td><code>created_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>projects.created_at</code></td><td>ms epoch</td></tr>
+        <tr class="f-stored"><td><code>updated_at</code></td><td><span class="badge" style="background:#1a3a2a;color:var(--green)">STORED</span></td><td><code>projects.updated_at</code></td><td>ms epoch</td></tr>
+        <tr class="f-ignored"><td><code>name</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>description</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>visibility</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-ignored"><td><code>archived</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>&mdash;</td><td>Deserialized but not stored</td></tr>
+        <tr class="f-dropped"><td><code>forks_count</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td></td></tr>
+        <tr class="f-dropped"><td><code>star_count</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td></td></tr>
+        <tr class="f-dropped"><td><code>issues_enabled</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Feature toggle</td></tr>
+        <tr class="f-dropped"><td><code>merge_requests_enabled</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Feature toggle</td></tr>
+        <tr class="f-dropped"><td><code>wiki_enabled</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td></td></tr>
+        <tr class="f-dropped"><td><code>http_url_to_repo</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Clone URL</td></tr>
+        <tr class="f-dropped"><td><code>ssh_url_to_repo</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Clone URL</td></tr>
+        <tr class="f-dropped"><td><code>owner</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Owner user object</td></tr>
+        <tr class="f-dropped"><td><code>namespace</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Group/namespace info</td></tr>
+        <tr class="f-dropped"><td><code>last_activity_at</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Last push/update</td></tr>
+        <tr class="f-dropped"><td><code>statistics</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td>&mdash;</td><td>Commit count, storage, etc.</td></tr>
+      </table>
+    </div>
+  </div>
+
+  <!-- USER RESPONSE -->
+  <div class="expandable" data-id="fld-user">
+    <div class="expandable-header" onclick="toggleExpand('fld-user')">
+      <span><strong>GET /user</strong> &mdash; Current user response <span style="color:var(--text-muted)">(0 stored, transient only / 20+ total)</span></span>
+      <span class="arrow">&#9654;</span>
+    </div>
+    <div class="expandable-body" id="fld-user-body">
+      <table class="field-table">
+        <tr><th style="width:200px">API Field</th><th style="width:90px">Status</th><th>Notes</th></tr>
+        <tr class="f-transient"><td><code>id</code></td><td><span class="badge" style="background:#2a2a1a;color:var(--yellow)">TRANSIENT</span></td><td>Displayed in auth check</td></tr>
+        <tr class="f-transient"><td><code>username</code></td><td><span class="badge" style="background:#2a2a1a;color:var(--yellow)">TRANSIENT</span></td><td>Displayed in auth check</td></tr>
+        <tr class="f-transient"><td><code>name</code></td><td><span class="badge" style="background:#2a2a1a;color:var(--yellow)">TRANSIENT</span></td><td>Displayed in auth check</td></tr>
+        <tr class="f-ignored"><td><code>email</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>Deserialized but not used</td></tr>
+        <tr class="f-ignored"><td><code>avatar_url</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>Deserialized but not used</td></tr>
+        <tr class="f-ignored"><td><code>web_url</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>Deserialized but not used</td></tr>
+        <tr class="f-ignored"><td><code>created_at</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>Deserialized but not used</td></tr>
+        <tr class="f-ignored"><td><code>state</code></td><td><span class="badge" style="background:#2a1a2a;color:var(--purple)">IGNORED</span></td><td>Deserialized but not used</td></tr>
+        <tr class="f-dropped"><td><code>bio</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+        <tr class="f-dropped"><td><code>location</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+        <tr class="f-dropped"><td><code>public_email</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+        <tr class="f-dropped"><td><code>organization</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+        <tr class="f-dropped"><td><code>job_title</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+        <tr class="f-dropped"><td><code>two_factor_enabled</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+        <tr class="f-dropped"><td><code>identities</code></td><td><span class="badge" style="background:#2a1a1a;color:var(--red)">DROPPED</span></td><td></td></tr>
+      </table>
+    </div>
+  </div>
+
+  <!-- SUMMARY STATS -->
+  <h3 style="margin-top:24px">Field Coverage Summary</h3>
+  <div class="grid-2">
+    <div class="card">
+      <div class="card-title">Issues Response</div>
+      <div class="stat-row"><span style="color:var(--green)">Stored</span><span>15 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:36%;background:var(--green)"></div></div>
+      <div class="stat-row"><span style="color:var(--purple)">Deserialized, ignored</span><span>5 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:12%;background:var(--purple)"></div></div>
+      <div class="stat-row"><span style="color:var(--red)">Never deserialized</span><span>~22 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:52%;background:var(--red)"></div></div>
+    </div>
+    <div class="card">
+      <div class="card-title">Merge Request Response</div>
+      <div class="stat-row"><span style="color:var(--green)">Stored</span><span>23 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:48%;background:var(--green)"></div></div>
+      <div class="stat-row"><span style="color:var(--yellow)">Transient (fallbacks)</span><span>3 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:6%;background:var(--yellow)"></div></div>
+      <div class="stat-row"><span style="color:var(--red)">Never deserialized</span><span>~22 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:46%;background:var(--red)"></div></div>
+    </div>
+    <div class="card">
+      <div class="card-title">Discussion/Note Response</div>
+      <div class="stat-row"><span style="color:var(--green)">Stored</span><span>23 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:64%;background:var(--green)"></div></div>
+      <div class="stat-row"><span style="color:var(--red)">Never deserialized</span><span>~13 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:36%;background:var(--red)"></div></div>
+    </div>
+    <div class="card">
+      <div class="card-title">Project Response</div>
+      <div class="stat-row"><span style="color:var(--green)">Stored</span><span>6 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:14%;background:var(--green)"></div></div>
+      <div class="stat-row"><span style="color:var(--purple)">Deserialized, ignored</span><span>4 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:10%;background:var(--purple)"></div></div>
+      <div class="stat-row"><span style="color:var(--red)">Never deserialized</span><span>~30 fields</span></div>
+      <div class="coverage-bar"><div class="coverage-fill" style="width:76%;background:var(--red)"></div></div>
+    </div>
+  </div>
+
+  <div class="card" style="margin-top:16px;border-color:var(--accent)">
+    <div class="card-title">Key insight: Raw payloads preserve everything</div>
+    <p style="font-size:13px;color:var(--text-muted)">Although many fields are dropped during transformation, the <code>raw_payloads</code> table stores the complete original JSON response (with SHA-256 dedup and optional gzip). This means all "dropped" data is still recoverable from the blob storage without re-fetching from GitLab. The normalized tables are optimized for query patterns, not completeness.</p>
+  </div>
+</div>
+
+<!-- EFFICIENCY ANALYSIS -->
+<div id="efficiency" class="panel">
+  <h2>Efficiency Analysis &amp; Opportunities</h2>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:16px">Observations on how gitlore could leverage the GitLab API more efficiently, and data it currently leaves on the table.</p>
+
+  <h3>Current Efficiency Wins</h3>
+  <div class="grid-2">
+    <div class="card" style="border-left:3px solid var(--green)">
+      <div class="card-title" style="color:var(--green)">Cursor-based incremental sync</div>
+      <p style="font-size:13px;color:var(--text-muted)">Uses <code>updated_after</code> + <code>order_by=updated_at&amp;sort=asc</code> to only fetch changed records. Avoids full re-fetch on every sync. This is the single biggest efficiency feature.</p>
+    </div>
+    <div class="card" style="border-left:3px solid var(--green)">
+      <div class="card-title" style="color:var(--green)">Raw payload dedup</div>
+      <p style="font-size:13px;color:var(--text-muted)">SHA-256 hashing prevents storing identical payloads. If an issue's <code>updated_at</code> changes but the actual content is identical, the raw blob is deduplicated.</p>
+    </div>
+    <div class="card" style="border-left:3px solid var(--green)">
+      <div class="card-title" style="color:var(--green)">Discussion watermark</div>
+      <p style="font-size:13px;color:var(--text-muted)">Only re-syncs discussions for issues/MRs whose <code>updated_at</code> has advanced past their <code>discussions_synced_for_updated_at</code> watermark. Skips unchanged entities.</p>
+    </div>
+    <div class="card" style="border-left:3px solid var(--green)">
+      <div class="card-title" style="color:var(--green)">Parallel discussion prefetch</div>
+      <p style="font-size:13px;color:var(--text-muted)">Fetches discussions for multiple issues/MRs concurrently (configurable, default 2). Dramatically reduces wall-clock time for discussion sync.</p>
+    </div>
+  </div>
+
+  <h3>Potential Inefficiencies</h3>
+
+  <div class="card" style="border-left:3px solid var(--red)">
+    <div class="card-title" style="color:var(--red)">1. Discussion full-refresh strategy</div>
+    <p style="font-size:13px">Every time an issue/MR is updated, ALL its discussions are re-fetched and replaced (DELETE + INSERT). For heavily-discussed items (50+ comments), this is expensive.</p>
+    <table style="margin-top:8px">
+      <tr><th>Scenario</th><th>Current</th><th>Alternative</th></tr>
+      <tr><td>Issue with 100 notes gets 1 new comment</td><td>Re-fetch all 100 notes (multiple pages)</td><td>Could use <code>GET .../notes?order_by=updated_at&amp;updated_after=...</code> for incremental note sync</td></tr>
+      <tr><td>MR label change (no new comments)</td><td>Re-fetch all discussions anyway</td><td>Could check <code>user_notes_count</code> delta or use Notes API with updated_after</td></tr>
+    </table>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px"><strong>Trade-off:</strong> Full-refresh is simpler and guarantees consistency (catches edits, deletes). Incremental would miss deleted notes.</p>
+  </div>
+
+  <div class="card" style="border-left:3px solid var(--yellow)">
+    <div class="card-title" style="color:var(--yellow)">2. Offset pagination instead of keyset</div>
+    <p style="font-size:13px">Gitlore uses <code>page=N&amp;per_page=100</code> offset pagination. GitLab supports <strong>keyset pagination</strong> for some endpoints (Issues, MRs), which is more efficient for large datasets and recommended by GitLab.</p>
+    <pre style="margin-top:8px">Current:  GET /projects/:id/issues?page=5&per_page=100
+Keyset:   GET /projects/:id/issues?pagination=keyset&per_page=100
+          (uses Link header rel="next" with cursor)</pre>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px"><strong>Benefit:</strong> Keyset pagination is O(1) per page (vs O(N) for offset). GitLab recommends it for &gt;10,000 records. Gitlore already parses <code>Link</code> headers, so the client-side support partially exists.</p>
+  </div>
+
+  <div class="card" style="border-left:3px solid var(--yellow)">
+    <div class="card-title" style="color:var(--yellow)">3. No ETag / conditional request support</div>
+    <p style="font-size:13px">GitLab returns <code>ETag</code> headers on API responses. Sending <code>If-None-Match</code> on subsequent requests would return <code>304 Not Modified</code> without consuming rate limit quota on some endpoints. Currently all requests are unconditional.</p>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px"><strong>Impact:</strong> Moderate. The cursor-based sync already avoids re-fetching unchanged data, so ETag would mainly help with the <code>discussions</code> full-refresh scenario where nothing changed.</p>
+  </div>
+
+  <div class="card" style="border-left:3px solid var(--yellow)">
+    <div class="card-title" style="color:var(--yellow)">4. Labels extracted from embedded data, not dedicated API</div>
+    <p style="font-size:13px">Gitlore extracts labels from the <code>labels[]</code> string array embedded in issue/MR responses. The dedicated <code>GET /projects/:id/labels</code> endpoint returns richer data:</p>
+    <table style="margin-top:8px">
+      <tr><th>From issues response</th><th>From Labels API</th></tr>
+      <tr><td>Label name (string only)</td><td>name, color, description, text_color, priority, is_project_label, subscribed, open_issues_count, closed_issues_count, open_merge_requests_count</td></tr>
+    </table>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px"><strong>Impact:</strong> The <code>labels</code> table has <code>color</code> and <code>description</code> columns but they may not be populated from the embedded string array. A single Labels API call (one request, non-paginated for most projects) would enrich the local label catalog.</p>
+  </div>
+
+  <h3>Dropped Data Worth Capturing</h3>
+  <p style="font-size:13px;color:var(--text-muted);margin-bottom:12px">Fields currently silently dropped that could add value to local queries:</p>
+
+  <table>
+    <tr><th>Field</th><th>Source</th><th>Value Proposition</th><th>Effort</th></tr>
+    <tr>
+      <td><code>user_notes_count</code></td>
+      <td>Issues, MRs</td>
+      <td>Could skip discussion re-sync when count hasn't changed. Quick "activity" sort without joining notes table.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>upvotes</code> / <code>downvotes</code></td>
+      <td>Issues, MRs</td>
+      <td>Engagement metrics for triage. "Most upvoted issues" is a common query.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>confidential</code></td>
+      <td>Issues</td>
+      <td>Security-sensitive filtering. Avoid exposing confidential issues in outputs.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>weight</code></td>
+      <td>Issues</td>
+      <td>Effort estimation for sprint planning (Premium/Ultimate only).</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>time_stats</code></td>
+      <td>Issues, MRs</td>
+      <td>Time tracking data for project reporting. Already in the response, free to capture.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>has_conflicts</code></td>
+      <td>MRs</td>
+      <td>Identify MRs needing rebase. Useful for "stale MR" alerts.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>blocking_discussions_resolved</code></td>
+      <td>MRs</td>
+      <td>MR readiness indicator without joining discussions table.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>merge_commit_sha</code></td>
+      <td>MRs</td>
+      <td>Trace merged MRs to specific commits. Useful for git correlation.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>suggestions[]</code></td>
+      <td>Discussion notes</td>
+      <td>Code review suggestions with from/to content. Rich data for code review analysis.</td>
+      <td style="color:var(--yellow)">Medium</td>
+    </tr>
+    <tr>
+      <td><code>task_completion_status</code></td>
+      <td>Issues, MRs</td>
+      <td>Track task-list checkbox progress without parsing description markdown.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>issue_type</code></td>
+      <td>Issues</td>
+      <td>Distinguish issues vs incidents vs test cases.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+    <tr>
+      <td><code>discussion_locked</code></td>
+      <td>Issues, MRs</td>
+      <td>Know if new comments can be added.</td>
+      <td style="color:var(--green)">Low</td>
+    </tr>
+  </table>
+
+  <h3>Structural Optimization Opportunities</h3>
+
+  <div class="card" style="border-left:3px solid var(--purple)">
+    <div class="card-title" style="color:var(--purple)">5. User denormalization</div>
+    <p style="font-size:13px">Currently stores only <code>username</code> for authors, assignees, reviewers. The API returns <code>name</code>, <code>avatar_url</code>, <code>web_url</code>, and <code>state</code> for every user reference. A <code>users</code> table could deduplicate this data and provide richer displays.</p>
+    <pre style="margin-top:8px">-- Potential schema
+CREATE TABLE users (
+  username TEXT PRIMARY KEY,
+  name TEXT,
+  gitlab_id INTEGER,
+  avatar_url TEXT,
+  state TEXT,           -- "active", "blocked", etc.
+  last_seen_at INTEGER  -- auto-updated on encounter
+);</pre>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px"><strong>Cost:</strong> No additional API calls. Data is already in every issue/MR/note response. Just needs extraction during transform.</p>
+  </div>
+
+  <div class="card" style="border-left:3px solid var(--purple)">
+    <div class="card-title" style="color:var(--purple)">6. MR milestone not captured</div>
+    <p style="font-size:13px">Milestones are stored for issues but the MR transformer does not extract the <code>milestone</code> object from MR responses, even though GitLab returns it. The <code>merge_requests</code> table has no <code>milestone_id</code> column.</p>
+    <p style="font-size:12px;color:var(--text-muted);margin-top:8px"><strong>Impact:</strong> Cannot query "which MRs are in milestone X?" locally. The data is in the raw payload but not indexed.</p>
+  </div>
+
+  <div class="card" style="border-left:3px solid var(--purple)">
+    <div class="card-title" style="color:var(--purple)">7. Issue references not captured</div>
+    <p style="font-size:13px">MRs store <code>references.short</code> and <code>references.full</code>, but the issue transformer drops the <code>references</code> object entirely. This means issues lack the cross-project reference format (e.g., <code>group/project#42</code>).</p>
+  </div>
+
+  <h3>API Strategies Not Yet Used</h3>
+
+  <div class="card">
+    <div class="card-title">Webhooks (push-based sync)</div>
+    <p style="font-size:13px;color:var(--text-muted)">Instead of polling, GitLab can push events via <code>POST /projects/:id/hooks</code>. Would enable near-real-time sync without rate-limit cost. Requires a listener endpoint.</p>
+  </div>
+
+  <div class="card">
+    <div class="card-title">Events API (lightweight change detection)</div>
+    <p style="font-size:13px;color:var(--text-muted)"><code>GET /projects/:id/events</code> returns a stream of all project activity. Could be used as a fast "has anything changed?" check before running expensive issue/MR sync. Much lighter than fetching full issue lists.</p>
+  </div>
+
+  <div class="card">
+    <div class="card-title">GraphQL API (precise field selection)</div>
+    <p style="font-size:13px;color:var(--text-muted)">GitLab's GraphQL API allows requesting exactly the fields needed. Would eliminate bandwidth waste from ~50% of response fields being silently dropped. Trade-off: different pagination model, potentially less stable API surface.</p>
+  </div>
+
+  <h3 style="margin-top:24px">Summary Verdict</h3>
+  <div class="card" style="border:1px solid var(--accent)">
+    <p style="font-size:13px">Gitlore is <strong>well-optimized for its core use case</strong> (read-only local sync). The cursor-based incremental sync and raw payload archival are sophisticated. The main opportunities are:</p>
+    <ol style="font-size:13px;padding-left:20px;margin-top:8px">
+      <li><strong>Capture more "free" data</strong> &mdash; Fields like <code>user_notes_count</code>, <code>upvotes</code>, <code>has_conflicts</code> are already in API responses. Storing them costs zero API calls and enables richer queries.</li>
+      <li><strong>Discussion sync efficiency</strong> &mdash; The full-refresh strategy is the biggest source of redundant API calls. Even a simple <code>user_notes_count</code> comparison could skip unchanged discussions.</li>
+      <li><strong>Keyset pagination</strong> &mdash; A meaningful improvement for large projects (&gt;10K issues), and Gitlore already has partial infrastructure for it.</li>
+      <li><strong>MR milestone parity</strong> &mdash; Low-effort gap to close with issue milestone support.</li>
+    </ol>
+  </div>
+</div>
+
+</div><!-- /content -->
+
+<script>
+function showTab(id) {
+  document.querySelectorAll('.tab').forEach(t => t.classList.remove('active'));
+  document.querySelectorAll('.panel').forEach(p => p.classList.remove('active'));
+  document.getElementById(id).classList.add('active');
+  document.querySelector(`.tab[onclick="showTab('${id}')"]`).classList.add('active');
+}
+
+function toggleExpand(id) {
+  const header = document.querySelector(`[data-id="${id}"] .expandable-header`);
+  const body = document.getElementById(`${id}-body`);
+  header.classList.toggle('open');
+  body.classList.toggle('open');
+}
+
+function filterAPI() {
+  const query = document.getElementById('api-search').value.toLowerCase();
+  document.querySelectorAll('#api-list .api-table tr').forEach(row => {
+    if (row.querySelector('th')) return; // skip headers
+    const text = row.textContent.toLowerCase();
+    row.style.display = text.includes(query) ? '' : 'none';
+  });
+}
+
+// Style used rows
+document.querySelectorAll('.used-row').forEach(row => {
+  row.style.background = 'rgba(88, 166, 255, 0.05)';
+});
+
+function filterFields() {
+  const query = document.getElementById('field-search').value.toLowerCase();
+  document.querySelectorAll('#fields .field-table tr').forEach(row => {
+    if (row.querySelector('th')) return;
+    const text = row.textContent.toLowerCase();
+    row.style.display = text.includes(query) ? '' : 'none';
+  });
+}
+</script>
+
+</body>
+</html>
diff --git a/phase-a-review.html b/phase-a-review.html
new file mode 100644
index 0000000..5baf68e
--- /dev/null
+++ b/phase-a-review.html
@@ -0,0 +1,1260 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>Phase A: Complete API Field Capture — Review</title>
+<style>
+  :root {
+    --bg: #0d1117;
+    --surface: #161b22;
+    --surface2: #1c2129;
+    --border: #30363d;
+    --text: #e6edf3;
+    --text-dim: #8b949e;
+    --accent: #58a6ff;
+    --green: #3fb950;
+    --green-bg: #0d2818;
+    --yellow: #d29922;
+    --yellow-bg: #2d2000;
+    --red: #f85149;
+    --red-bg: #3d1418;
+    --purple: #bc8cff;
+    --purple-bg: #1e1534;
+    --blue-bg: #0c2d6b;
+    --orange: #db6d28;
+    --cyan: #39d2c0;
+  }
+  * { margin: 0; padding: 0; box-sizing: border-box; }
+  body {
+    font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Helvetica, Arial, sans-serif;
+    background: var(--bg);
+    color: var(--text);
+    line-height: 1.6;
+    min-height: 100vh;
+  }
+
+  /* Header */
+  .header {
+    background: var(--surface);
+    border-bottom: 1px solid var(--border);
+    padding: 20px 32px;
+    position: sticky;
+    top: 0;
+    z-index: 100;
+  }
+  .header h1 { font-size: 20px; font-weight: 600; }
+  .header .subtitle { color: var(--text-dim); font-size: 13px; margin-top: 2px; }
+  .header .badges { display: flex; gap: 8px; margin-top: 10px; flex-wrap: wrap; }
+  .badge {
+    display: inline-flex; align-items: center; gap: 5px;
+    padding: 3px 10px; border-radius: 12px; font-size: 12px; font-weight: 500;
+  }
+  .badge-green { background: var(--green-bg); color: var(--green); border: 1px solid #1a4d2e; }
+  .badge-yellow { background: var(--yellow-bg); color: var(--yellow); border: 1px solid #4d3800; }
+  .badge-purple { background: var(--purple-bg); color: var(--purple); border: 1px solid #3b2a5e; }
+  .badge-blue { background: var(--blue-bg); color: var(--accent); border: 1px solid #1a4480; }
+
+  /* Tabs */
+  .tab-bar {
+    display: flex;
+    background: var(--surface);
+    border-bottom: 1px solid var(--border);
+    padding: 0 32px;
+    overflow-x: auto;
+    position: sticky;
+    top: 72px;
+    z-index: 99;
+  }
+  .tab {
+    padding: 10px 16px;
+    font-size: 13px;
+    color: var(--text-dim);
+    cursor: pointer;
+    border-bottom: 2px solid transparent;
+    white-space: nowrap;
+    transition: color 0.15s, border-color 0.15s;
+    user-select: none;
+  }
+  .tab:hover { color: var(--text); }
+  .tab.active { color: var(--accent); border-bottom-color: var(--accent); }
+  .tab .count {
+    background: var(--surface2);
+    color: var(--text-dim);
+    font-size: 11px;
+    padding: 1px 6px;
+    border-radius: 8px;
+    margin-left: 5px;
+  }
+
+  /* Content */
+  .content { max-width: 1100px; margin: 0 auto; padding: 24px 32px 80px; }
+  .tab-content { display: none; }
+  .tab-content.active { display: block; }
+
+  /* Cards */
+  .card {
+    background: var(--surface);
+    border: 1px solid var(--border);
+    border-radius: 8px;
+    margin-bottom: 16px;
+    overflow: hidden;
+  }
+  .card-header {
+    padding: 14px 18px;
+    border-bottom: 1px solid var(--border);
+    font-weight: 600;
+    font-size: 14px;
+    display: flex;
+    align-items: center;
+    justify-content: space-between;
+  }
+  .card-body { padding: 18px; }
+
+  /* Principle banner */
+  .principle {
+    background: var(--blue-bg);
+    border: 1px solid #1a4480;
+    border-radius: 8px;
+    padding: 16px 20px;
+    margin-bottom: 20px;
+    font-size: 14px;
+  }
+  .principle strong { color: var(--accent); }
+
+  /* Scope cards */
+  .scope-grid { display: grid; grid-template-columns: repeat(auto-fit, minmax(220px, 1fr)); gap: 12px; margin-bottom: 20px; }
+  .scope-item {
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 6px;
+    padding: 14px;
+  }
+  .scope-item .num { font-size: 28px; font-weight: 700; color: var(--accent); }
+  .scope-item .label { font-size: 12px; color: var(--text-dim); margin-top: 2px; }
+
+  /* Exclusion cards */
+  .exclusion {
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 6px;
+    padding: 14px 18px;
+    margin-bottom: 10px;
+  }
+  .exclusion .field-name { font-weight: 600; color: var(--red); font-size: 14px; }
+  .exclusion .rationale { color: var(--text-dim); font-size: 13px; margin-top: 4px; }
+  .exclusion .note { color: var(--yellow); font-size: 12px; margin-top: 6px; font-style: italic; }
+
+  /* Tables */
+  table { width: 100%; border-collapse: collapse; font-size: 13px; }
+  th {
+    text-align: left;
+    padding: 10px 14px;
+    background: var(--surface2);
+    color: var(--text-dim);
+    font-weight: 600;
+    font-size: 11px;
+    text-transform: uppercase;
+    letter-spacing: 0.5px;
+    border-bottom: 1px solid var(--border);
+    position: sticky;
+    top: 0;
+    z-index: 1;
+  }
+  td {
+    padding: 8px 14px;
+    border-bottom: 1px solid var(--border);
+    vertical-align: top;
+  }
+  tr:hover td { background: rgba(88,166,255,0.04); }
+  code {
+    font-family: 'SF Mono', 'Fira Code', 'Consolas', monospace;
+    font-size: 12px;
+    background: var(--surface2);
+    padding: 2px 6px;
+    border-radius: 4px;
+    color: var(--cyan);
+  }
+  td code { background: transparent; padding: 0; }
+
+  /* Status dots */
+  .dot {
+    display: inline-block;
+    width: 8px;
+    height: 8px;
+    border-radius: 50%;
+    margin-right: 6px;
+    position: relative;
+    top: -1px;
+  }
+  .dot-green { background: var(--green); }
+  .dot-yellow { background: var(--yellow); }
+  .dot-purple { background: var(--purple); }
+  .dot-red { background: var(--red); }
+
+  /* Category badges in tables */
+  .cat {
+    display: inline-block;
+    padding: 2px 8px;
+    border-radius: 10px;
+    font-size: 11px;
+    font-weight: 500;
+    white-space: nowrap;
+  }
+  .cat-meta { background: #1a2740; color: #79b8ff; }
+  .cat-engagement { background: #2a1e00; color: #f0b832; }
+  .cat-time { background: #1e2d1e; color: #7ee787; }
+  .cat-task { background: #2d1e3a; color: #d2a8ff; }
+  .cat-ref { background: #1a3040; color: #56d4dd; }
+  .cat-tracking { background: #2d1a1a; color: #ff9492; }
+  .cat-premium { background: #2a1a3d; color: #bc8cff; }
+  .cat-merge { background: #0d2818; color: #3fb950; }
+  .cat-fork { background: #1e2d1e; color: #7ee787; }
+  .cat-branch { background: #2d2700; color: #e3b341; }
+  .cat-schedule { background: #1a2740; color: #79b8ff; }
+  .cat-import { background: #2d1a1a; color: #ff9492; }
+  .cat-author { background: #1a3040; color: #56d4dd; }
+
+  /* Filter bar */
+  .filter-bar {
+    display: flex;
+    gap: 8px;
+    padding: 12px 0;
+    flex-wrap: wrap;
+    align-items: center;
+  }
+  .filter-btn {
+    padding: 4px 12px;
+    border-radius: 14px;
+    font-size: 12px;
+    border: 1px solid var(--border);
+    background: var(--surface);
+    color: var(--text-dim);
+    cursor: pointer;
+    transition: all 0.15s;
+    user-select: none;
+  }
+  .filter-btn:hover { border-color: var(--accent); color: var(--text); }
+  .filter-btn.active { background: var(--blue-bg); border-color: var(--accent); color: var(--accent); }
+  .search-input {
+    flex: 1;
+    min-width: 180px;
+    padding: 6px 12px;
+    border-radius: 6px;
+    border: 1px solid var(--border);
+    background: var(--surface2);
+    color: var(--text);
+    font-size: 13px;
+    outline: none;
+  }
+  .search-input:focus { border-color: var(--accent); }
+  .search-input::placeholder { color: var(--text-dim); }
+
+  /* SQL block */
+  .sql-block {
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 8px;
+    padding: 18px;
+    font-family: 'SF Mono', 'Fira Code', 'Consolas', monospace;
+    font-size: 12.5px;
+    line-height: 1.7;
+    overflow-x: auto;
+    white-space: pre;
+    color: var(--text);
+    max-height: 600px;
+    overflow-y: auto;
+  }
+  .sql-comment { color: var(--text-dim); }
+  .sql-keyword { color: var(--accent); font-weight: 600; }
+  .sql-type { color: var(--green); }
+  .sql-default { color: var(--yellow); }
+  .sql-divider { color: var(--text-dim); }
+  .sql-line { display: block; padding: 0 4px; }
+  .sql-line:hover { background: rgba(88,166,255,0.06); }
+
+  /* Struct block */
+  .struct-block {
+    background: var(--surface2);
+    border: 1px solid var(--border);
+    border-radius: 8px;
+    padding: 18px;
+    font-family: 'SF Mono', 'Fira Code', 'Consolas', monospace;
+    font-size: 12.5px;
+    line-height: 1.8;
+    overflow-x: auto;
+  }
+  .struct-field { color: var(--text); }
+  .struct-type { color: var(--green); }
+  .struct-comment { color: var(--text-dim); }
+  .struct-line { display: block; padding: 0 4px; }
+  .struct-line:hover { background: rgba(88,166,255,0.06); }
+
+  /* File list */
+  .file-item {
+    display: flex;
+    align-items: flex-start;
+    gap: 12px;
+    padding: 12px 16px;
+    border-bottom: 1px solid var(--border);
+  }
+  .file-item:last-child { border-bottom: none; }
+  .file-icon {
+    width: 32px; height: 32px;
+    border-radius: 6px;
+    display: flex;
+    align-items: center;
+    justify-content: center;
+    font-size: 14px;
+    flex-shrink: 0;
+  }
+  .file-icon-new { background: var(--green-bg); color: var(--green); }
+  .file-icon-mod { background: var(--yellow-bg); color: var(--yellow); }
+  .file-path { font-family: 'SF Mono', monospace; font-size: 13px; color: var(--accent); }
+  .file-desc { font-size: 12px; color: var(--text-dim); margin-top: 2px; }
+
+  /* Decision table */
+  .decision-row {
+    display: flex;
+    gap: 16px;
+    padding: 14px 18px;
+    border-bottom: 1px solid var(--border);
+    align-items: flex-start;
+  }
+  .decision-row:last-child { border-bottom: none; }
+  .decision-field {
+    font-weight: 600;
+    min-width: 100px;
+    font-family: 'SF Mono', monospace;
+    font-size: 13px;
+  }
+  .decision-verdict {
+    font-weight: 600;
+    min-width: 140px;
+  }
+  .verdict-excluded { color: var(--red); }
+  .verdict-flatten { color: var(--green); }
+  .decision-rationale { color: var(--text-dim); font-size: 13px; flex: 1; }
+
+  /* Transform rules */
+  .transform-rule {
+    display: flex;
+    gap: 12px;
+    padding: 10px 16px;
+    border-bottom: 1px solid var(--border);
+    align-items: baseline;
+  }
+  .transform-rule:last-child { border-bottom: none; }
+  .transform-from {
+    font-family: 'SF Mono', monospace;
+    font-size: 12px;
+    color: var(--yellow);
+    min-width: 200px;
+  }
+  .transform-arrow { color: var(--text-dim); min-width: 24px; text-align: center; }
+  .transform-to { font-size: 13px; color: var(--text); flex: 1; }
+
+  /* Not-included list */
+  .not-included {
+    list-style: none;
+    padding: 0;
+  }
+  .not-included li {
+    padding: 8px 16px;
+    border-bottom: 1px solid var(--border);
+    font-size: 13px;
+    color: var(--text-dim);
+  }
+  .not-included li:last-child { border-bottom: none; }
+  .not-included li::before {
+    content: "\2717";
+    color: var(--red);
+    margin-right: 10px;
+    font-weight: bold;
+  }
+
+  /* Column count summary */
+  .col-summary {
+    display: grid;
+    grid-template-columns: 1fr 1fr;
+    gap: 16px;
+    margin-bottom: 20px;
+  }
+  .col-summary-card {
+    background: var(--surface);
+    border: 1px solid var(--border);
+    border-radius: 8px;
+    padding: 18px;
+  }
+  .col-summary-card h3 { font-size: 14px; margin-bottom: 10px; }
+  .col-bar {
+    display: flex;
+    height: 8px;
+    border-radius: 4px;
+    overflow: hidden;
+    margin: 8px 0;
+    background: var(--surface2);
+  }
+  .col-bar-existing { background: var(--text-dim); }
+  .col-bar-new { background: var(--green); }
+  .col-legend {
+    display: flex;
+    gap: 16px;
+    font-size: 11px;
+    color: var(--text-dim);
+  }
+  .col-legend span { display: flex; align-items: center; gap: 4px; }
+  .legend-dot { width: 8px; height: 8px; border-radius: 2px; display: inline-block; }
+
+  /* Toggles for SQL sections */
+  .toggle-header {
+    cursor: pointer;
+    user-select: none;
+  }
+  .toggle-header::after {
+    content: " \25BC";
+    font-size: 10px;
+    color: var(--text-dim);
+  }
+  .toggle-header.collapsed::after {
+    content: " \25B6";
+  }
+
+  /* Responsive */
+  @media (max-width: 768px) {
+    .header { padding: 14px 16px; }
+    .tab-bar { padding: 0 16px; }
+    .content { padding: 16px; }
+    .col-summary { grid-template-columns: 1fr; }
+    .scope-grid { grid-template-columns: 1fr 1fr; }
+  }
+</style>
+</head>
+<body>
+
+<div class="header">
+  <h1>Phase A: Complete API Field Capture</h1>
+  <div class="subtitle">Migration 007 &mdash; Mirror all GitLab API response data into the local DB</div>
+  <div class="badges">
+    <span class="badge badge-green">No new API calls</span>
+    <span class="badge badge-green">No new tables</span>
+    <span class="badge badge-yellow">1 migration</span>
+    <span class="badge badge-purple">6 files touched</span>
+    <span class="badge badge-blue">Independent of CP3</span>
+  </div>
+</div>
+
+<div class="tab-bar" id="tabBar">
+  <div class="tab active" data-tab="overview">Overview</div>
+  <div class="tab" data-tab="issues">Issues <span class="count" id="issueCount"></span></div>
+  <div class="tab" data-tab="mrs">Merge Requests <span class="count" id="mrCount"></span></div>
+  <div class="tab" data-tab="migration">Migration SQL</div>
+  <div class="tab" data-tab="structs">Serde Structs</div>
+  <div class="tab" data-tab="transforms">Transformers</div>
+  <div class="tab" data-tab="files">Files Touched</div>
+  <div class="tab" data-tab="decisions">Decisions</div>
+</div>
+
+<div class="content">
+
+  <!-- OVERVIEW -->
+  <div class="tab-content active" id="tab-overview">
+    <div class="principle">
+      <strong>Guiding Principle:</strong> Mirror everything GitLab gives us. The local DB should be a complete representation of all data returned by the API. This ensures maximum context for processing and analysis in later steps.
+    </div>
+
+    <div class="scope-grid">
+      <div class="scope-item">
+        <div class="num" id="totalNewCols">0</div>
+        <div class="label">New columns total</div>
+      </div>
+      <div class="scope-item">
+        <div class="num" id="issueNewCols">0</div>
+        <div class="label">New issue columns</div>
+      </div>
+      <div class="scope-item">
+        <div class="num" id="mrNewCols">0</div>
+        <div class="label">New MR columns</div>
+      </div>
+      <div class="scope-item">
+        <div class="num">6</div>
+        <div class="label">New serde helper types</div>
+      </div>
+      <div class="scope-item">
+        <div class="num">2</div>
+        <div class="label">Fields excluded</div>
+      </div>
+      <div class="scope-item">
+        <div class="num">6</div>
+        <div class="label">Files touched</div>
+      </div>
+    </div>
+
+    <div class="col-summary">
+      <div class="col-summary-card">
+        <h3>Issues table</h3>
+        <div class="col-bar">
+          <div class="col-bar-existing" id="issueBarExisting"></div>
+          <div class="col-bar-new" id="issueBarNew"></div>
+        </div>
+        <div class="col-legend">
+          <span><span class="legend-dot" style="background:var(--text-dim)"></span> <span id="issueExistingNum">18</span> existing</span>
+          <span><span class="legend-dot" style="background:var(--green)"></span> <span id="issueNewNum">0</span> new</span>
+        </div>
+      </div>
+      <div class="col-summary-card">
+        <h3>Merge Requests table</h3>
+        <div class="col-bar">
+          <div class="col-bar-existing" id="mrBarExisting"></div>
+          <div class="col-bar-new" id="mrBarNew"></div>
+        </div>
+        <div class="col-legend">
+          <span><span class="legend-dot" style="background:var(--text-dim)"></span> <span id="mrExistingNum">29</span> existing</span>
+          <span><span class="legend-dot" style="background:var(--green)"></span> <span id="mrNewNum">0</span> new</span>
+        </div>
+      </div>
+    </div>
+
+    <div class="card">
+      <div class="card-header">Scope</div>
+      <div class="card-body" style="font-size:14px;">
+        <p style="margin-bottom:10px;">One migration. Three categories of work:</p>
+        <ol style="padding-left:20px; color:var(--text-dim); line-height:2;">
+          <li><strong style="color:var(--text)">New columns</strong> on <code>issues</code> and <code>merge_requests</code> for fields currently dropped by serde or during transform</li>
+          <li><strong style="color:var(--text)">New serde fields</strong> on <code>GitLabIssue</code> and <code>GitLabMergeRequest</code> to deserialize currently-silently-dropped JSON fields</li>
+          <li><strong style="color:var(--text)">Transformer + insert updates</strong> to pass the new fields through to the DB</li>
+        </ol>
+      </div>
+    </div>
+
+    <div class="card">
+      <div class="card-header">What this does NOT include</div>
+      <div class="card-body" style="padding:0;">
+        <ul class="not-included">
+          <li>No new API endpoints called</li>
+          <li>No new tables (except reusing existing <code>milestones</code> for MRs)</li>
+          <li>No CLI changes (new fields stored but not surfaced in <code>lore issues</code> / <code>lore mrs</code>)</li>
+          <li>No changes to discussion/note ingestion (Phase A is issues + MRs only)</li>
+          <li>No observability instrumentation (that's Phase B)</li>
+        </ul>
+      </div>
+    </div>
+  </div>
+
+  <!-- ISSUES TAB -->
+  <div class="tab-content" id="tab-issues">
+    <h2 style="font-size:16px; margin-bottom:16px;">Issues: Field Gap Inventory</h2>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">Currently Stored</div>
+      <div class="card-body" style="font-size:13px; color:var(--text-dim); line-height:2;">
+        <span style="color:var(--text)">id</span>, <span style="color:var(--text)">iid</span>, <span style="color:var(--text)">project_id</span>, <span style="color:var(--text)">title</span>, <span style="color:var(--text)">description</span>, <span style="color:var(--text)">state</span>, <span style="color:var(--text)">author_username</span>, <span style="color:var(--text)">created_at</span>, <span style="color:var(--text)">updated_at</span>, <span style="color:var(--text)">web_url</span>, <span style="color:var(--text)">due_date</span>, <span style="color:var(--text)">milestone_id</span>, <span style="color:var(--text)">milestone_title</span>, <span style="color:var(--text)">raw_payload_id</span>, <span style="color:var(--text)">last_seen_at</span>, <span style="color:var(--text)">discussions_synced_for_updated_at</span>, <span style="color:var(--text)">labels</span> (junction), <span style="color:var(--text)">assignees</span> (junction)
+      </div>
+    </div>
+
+    <div class="filter-bar">
+      <input type="text" class="search-input" id="issueSearch" placeholder="Filter fields...">
+      <button class="filter-btn active" data-filter="all" data-table="issues">All</button>
+      <button class="filter-btn" data-filter="wired" data-table="issues">Deserialized, not stored</button>
+      <button class="filter-btn" data-filter="new" data-table="issues">New from API</button>
+    </div>
+
+    <div class="card">
+      <div class="card-body" style="padding:0; overflow-x:auto;">
+        <table id="issueTable">
+          <thead>
+            <tr>
+              <th>API Field</th>
+              <th>Type</th>
+              <th>DB Column</th>
+              <th>Category</th>
+              <th>Status</th>
+              <th>Notes</th>
+            </tr>
+          </thead>
+          <tbody></tbody>
+        </table>
+      </div>
+    </div>
+  </div>
+
+  <!-- MRs TAB -->
+  <div class="tab-content" id="tab-mrs">
+    <h2 style="font-size:16px; margin-bottom:16px;">Merge Requests: Field Gap Inventory</h2>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">Currently Stored</div>
+      <div class="card-body" style="font-size:13px; color:var(--text-dim); line-height:2;">
+        <span style="color:var(--text)">id</span>, <span style="color:var(--text)">iid</span>, <span style="color:var(--text)">project_id</span>, <span style="color:var(--text)">title</span>, <span style="color:var(--text)">description</span>, <span style="color:var(--text)">state</span>, <span style="color:var(--text)">draft</span>, <span style="color:var(--text)">author_username</span>, <span style="color:var(--text)">source_branch</span>, <span style="color:var(--text)">target_branch</span>, <span style="color:var(--text)">head_sha</span>, <span style="color:var(--text)">references_short</span>, <span style="color:var(--text)">references_full</span>, <span style="color:var(--text)">detailed_merge_status</span>, <span style="color:var(--text)">merge_user_username</span>, <span style="color:var(--text)">created_at</span>, <span style="color:var(--text)">updated_at</span>, <span style="color:var(--text)">merged_at</span>, <span style="color:var(--text)">closed_at</span>, <span style="color:var(--text)">last_seen_at</span>, <span style="color:var(--text)">web_url</span>, <span style="color:var(--text)">raw_payload_id</span>, <span style="color:var(--text)">discussions_synced_for_updated_at</span>, <span style="color:var(--text)">labels</span> (junction), <span style="color:var(--text)">assignees</span> (junction), <span style="color:var(--text)">reviewers</span> (junction)
+      </div>
+    </div>
+
+    <div class="filter-bar">
+      <input type="text" class="search-input" id="mrSearch" placeholder="Filter fields...">
+      <button class="filter-btn active" data-filter="all" data-table="mrs">All</button>
+      <button class="filter-btn" data-filter="wired" data-table="mrs">Deserialized, not stored</button>
+      <button class="filter-btn" data-filter="new" data-table="mrs">New from API</button>
+    </div>
+
+    <div class="card">
+      <div class="card-body" style="padding:0; overflow-x:auto;">
+        <table id="mrTable">
+          <thead>
+            <tr>
+              <th>API Field</th>
+              <th>Type</th>
+              <th>DB Column</th>
+              <th>Category</th>
+              <th>Status</th>
+              <th>Notes</th>
+            </tr>
+          </thead>
+          <tbody></tbody>
+        </table>
+      </div>
+    </div>
+  </div>
+
+  <!-- MIGRATION SQL TAB -->
+  <div class="tab-content" id="tab-migration">
+    <h2 style="font-size:16px; margin-bottom:16px;">Migration 007: complete_field_capture.sql</h2>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">
+        <span>Issues columns</span>
+        <span class="badge badge-green" id="issueSqlCount"></span>
+      </div>
+      <div class="card-body" style="padding:0;">
+        <div class="sql-block" id="issueSql"></div>
+      </div>
+    </div>
+
+    <div class="card">
+      <div class="card-header">
+        <span>Merge Requests columns</span>
+        <span class="badge badge-green" id="mrSqlCount"></span>
+      </div>
+      <div class="card-body" style="padding:0;">
+        <div class="sql-block" id="mrSql"></div>
+      </div>
+    </div>
+  </div>
+
+  <!-- SERDE STRUCTS TAB -->
+  <div class="tab-content" id="tab-structs">
+    <h2 style="font-size:16px; margin-bottom:16px;">Serde Struct Changes</h2>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">New Helper Types</div>
+      <div class="card-body" style="padding:0;">
+        <div class="struct-block" id="helperTypes"></div>
+      </div>
+    </div>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">
+        <span>GitLabIssue: new fields</span>
+        <span class="badge badge-yellow" id="issueFieldCount"></span>
+      </div>
+      <div class="card-body" style="padding:0;">
+        <div class="struct-block" id="issueStruct"></div>
+      </div>
+    </div>
+
+    <div class="card">
+      <div class="card-header">
+        <span>GitLabMergeRequest: new fields</span>
+        <span class="badge badge-yellow" id="mrFieldCount"></span>
+      </div>
+      <div class="card-body" style="padding:0;">
+        <div class="struct-block" id="mrStruct"></div>
+      </div>
+    </div>
+  </div>
+
+  <!-- TRANSFORMERS TAB -->
+  <div class="tab-content" id="tab-transforms">
+    <h2 style="font-size:16px; margin-bottom:16px;">Transformer Changes</h2>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">IssueRow: transform rules</div>
+      <div class="card-body" style="padding:0;">
+        <p style="padding:12px 16px; font-size:13px; color:var(--text-dim); border-bottom:1px solid var(--border);">
+          All new fields map 1:1 from the serde struct <strong style="color:var(--text)">except</strong> these special cases:
+        </p>
+        <div id="issueTransforms"></div>
+      </div>
+    </div>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">NormalizedMergeRequest: transform rules</div>
+      <div class="card-body" style="padding:0;">
+        <p style="padding:12px 16px; font-size:13px; color:var(--text-dim); border-bottom:1px solid var(--border);">
+          Same patterns as issues, plus:
+        </p>
+        <div id="mrTransforms"></div>
+      </div>
+    </div>
+
+    <div class="card">
+      <div class="card-header">Insert statement changes</div>
+      <div class="card-body" style="font-size:13px; color:var(--text-dim);">
+        Both <code>process_issue_in_transaction</code> and <code>process_mr_in_transaction</code> need their
+        <strong style="color:var(--text)">INSERT</strong> and <strong style="color:var(--text)">ON CONFLICT DO UPDATE</strong>
+        statements extended with all new columns. The ON CONFLICT clause should update all new fields on re-sync.
+      </div>
+    </div>
+  </div>
+
+  <!-- FILES TOUCHED TAB -->
+  <div class="tab-content" id="tab-files">
+    <h2 style="font-size:16px; margin-bottom:16px;">Files Touched</h2>
+    <div class="card">
+      <div class="card-body" style="padding:0;" id="filesList"></div>
+    </div>
+  </div>
+
+  <!-- DECISIONS TAB -->
+  <div class="tab-content" id="tab-decisions">
+    <h2 style="font-size:16px; margin-bottom:16px;">Resolved Decisions</h2>
+
+    <div class="card" style="margin-bottom:20px;">
+      <div class="card-header">Exclusions (2 fields)</div>
+      <div class="card-body">
+        <div class="exclusion">
+          <div class="field-name"><code>subscribed</code> &mdash; Excluded</div>
+          <div class="rationale">User-relative field. Reflects the token holder's subscription state, not a property of the entity itself. Changes meaning if the token is rotated to a different user. Not entity data.</div>
+        </div>
+        <div class="exclusion">
+          <div class="field-name"><code>_links</code> &mdash; Excluded</div>
+          <div class="rationale">HATEOAS API navigation metadata, not entity data. Every URL is deterministically constructable from <code>project_id</code> + <code>iid</code> + GitLab base URL.</div>
+          <div class="note">Note: <code>closed_as_duplicate_of</code> inside <code>_links</code> contains a real entity reference. Extracting that is deferred to a future phase.</div>
+        </div>
+      </div>
+    </div>
+
+    <div class="card">
+      <div class="card-header">Included with special handling</div>
+      <div class="card-body">
+        <div class="exclusion" style="margin-bottom:0;">
+          <div class="field-name" style="color:var(--green);"><code>epic</code> / <code>iteration</code> &mdash; Flatten to columns</div>
+          <div class="rationale">Same denormalization pattern as milestones. Epic gets 5 columns (<code>epic_id</code>, <code>epic_iid</code>, <code>epic_title</code>, <code>epic_url</code>, <code>epic_group_id</code>). Iteration gets 6 columns (<code>iteration_id</code>, <code>iteration_iid</code>, <code>iteration_title</code>, <code>iteration_state</code>, <code>iteration_start_date</code>, <code>iteration_due_date</code>). Both nullable (null on Free tier).</div>
+        </div>
+      </div>
+    </div>
+  </div>
+
+</div>
+
+<script>
+// ============================================================
+// DATA
+// ============================================================
+
+const issueFields = [
+  // Deserialized but not stored
+  { api: 'closed_at', type: 'String', col: 'closed_at', cat: 'tracking', catLabel: 'Tracking', status: 'wired', notes: 'Deserialized in serde struct, DB column exists (migration 002), but transformer never populates it' },
+  { api: 'author.id', type: 'i64', col: 'author_id', cat: 'author', catLabel: 'Author', status: 'wired', notes: 'Already deserialized, just not passed to DB' },
+  { api: 'author.name', type: 'String', col: 'author_name', cat: 'author', catLabel: 'Author', status: 'wired', notes: 'Already deserialized, just not passed to DB' },
+  // New from API
+  { api: 'type', type: 'String', col: 'issue_type', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: '"ISSUE", "INCIDENT", "TEST_CASE"' },
+  { api: 'upvotes', type: 'i64', col: 'upvotes', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: '' },
+  { api: 'downvotes', type: 'i64', col: 'downvotes', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: '' },
+  { api: 'user_notes_count', type: 'i64', col: 'user_notes_count', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: 'Useful for discussion sync optimization' },
+  { api: 'merge_requests_count', type: 'i64', col: 'merge_requests_count', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: 'Count of linked MRs' },
+  { api: 'confidential', type: 'bool', col: 'confidential', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: '0/1' },
+  { api: 'discussion_locked', type: 'bool', col: 'discussion_locked', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: '0/1' },
+  { api: 'weight', type: 'Option<i64>', col: 'weight', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Premium/Ultimate, null on Free' },
+  { api: 'time_stats.time_estimate', type: 'i64', col: 'time_estimate', cat: 'time', catLabel: 'Time', status: 'new', notes: 'Seconds' },
+  { api: 'time_stats.total_time_spent', type: 'i64', col: 'time_spent', cat: 'time', catLabel: 'Time', status: 'new', notes: 'Seconds' },
+  { api: 'time_stats.human_time_estimate', type: 'Option<String>', col: 'human_time_estimate', cat: 'time', catLabel: 'Time', status: 'new', notes: 'e.g. "3h 30m"' },
+  { api: 'time_stats.human_total_time_spent', type: 'Option<String>', col: 'human_time_spent', cat: 'time', catLabel: 'Time', status: 'new', notes: 'e.g. "1h 15m"' },
+  { api: 'task_completion_status.count', type: 'i64', col: 'task_count', cat: 'task', catLabel: 'Tasks', status: 'new', notes: 'Checkbox total' },
+  { api: 'task_completion_status.completed_count', type: 'i64', col: 'task_completed_count', cat: 'task', catLabel: 'Tasks', status: 'new', notes: 'Checkboxes checked' },
+  { api: 'has_tasks', type: 'bool', col: 'has_tasks', cat: 'task', catLabel: 'Tasks', status: 'new', notes: '0/1' },
+  { api: 'severity', type: 'Option<String>', col: 'severity', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Incident severity' },
+  { api: 'closed_by', type: 'Option<object>', col: 'closed_by_username', cat: 'tracking', catLabel: 'Tracking', status: 'new', notes: 'Username only (consistent with author pattern)' },
+  { api: 'imported', type: 'bool', col: 'imported', cat: 'import', catLabel: 'Import', status: 'new', notes: '0/1' },
+  { api: 'imported_from', type: 'Option<String>', col: 'imported_from', cat: 'import', catLabel: 'Import', status: 'new', notes: 'Import source' },
+  { api: 'moved_to_id', type: 'Option<i64>', col: 'moved_to_id', cat: 'tracking', catLabel: 'Tracking', status: 'new', notes: 'Target issue if moved' },
+  { api: 'references.short', type: 'String', col: 'references_short', cat: 'ref', catLabel: 'References', status: 'new', notes: 'e.g. "#42"' },
+  { api: 'references.relative', type: 'String', col: 'references_relative', cat: 'ref', catLabel: 'References', status: 'new', notes: 'e.g. "#42" or "group/proj#42"' },
+  { api: 'references.full', type: 'String', col: 'references_full', cat: 'ref', catLabel: 'References', status: 'new', notes: 'e.g. "group/project#42"' },
+  { api: 'health_status', type: 'Option<String>', col: 'health_status', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Ultimate only' },
+  { api: 'issue_type', type: 'Option<String>', col: '(same as type)', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: 'Alias, use whichever GitLab returns' },
+  { api: 'epic.id', type: 'Option<i64>', col: 'epic_id', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Premium/Ultimate, null on Free' },
+  { api: 'epic.iid', type: 'Option<i64>', col: 'epic_iid', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'epic.title', type: 'Option<String>', col: 'epic_title', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'epic.url', type: 'Option<String>', col: 'epic_url', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'epic.group_id', type: 'Option<i64>', col: 'epic_group_id', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'iteration.id', type: 'Option<i64>', col: 'iteration_id', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Premium/Ultimate, null on Free' },
+  { api: 'iteration.iid', type: 'Option<i64>', col: 'iteration_iid', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'iteration.title', type: 'Option<String>', col: 'iteration_title', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'iteration.state', type: 'Option<i64>', col: 'iteration_state', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '1=upcoming, 2=current, 3=closed' },
+  { api: 'iteration.start_date', type: 'Option<String>', col: 'iteration_start_date', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'ISO date' },
+  { api: 'iteration.due_date', type: 'Option<String>', col: 'iteration_due_date', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'ISO date' },
+];
+
+const mrFields = [
+  // Deserialized but not stored
+  { api: 'author.id', type: 'i64', col: 'author_id', cat: 'author', catLabel: 'Author', status: 'wired', notes: 'Already deserialized, just not passed to DB' },
+  { api: 'author.name', type: 'String', col: 'author_name', cat: 'author', catLabel: 'Author', status: 'wired', notes: 'Already deserialized, just not passed to DB' },
+  // New from API
+  { api: 'upvotes', type: 'i64', col: 'upvotes', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: '' },
+  { api: 'downvotes', type: 'i64', col: 'downvotes', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: '' },
+  { api: 'user_notes_count', type: 'i64', col: 'user_notes_count', cat: 'engagement', catLabel: 'Engagement', status: 'new', notes: '' },
+  { api: 'source_project_id', type: 'i64', col: 'source_project_id', cat: 'fork', catLabel: 'Fork', status: 'new', notes: 'Fork source' },
+  { api: 'target_project_id', type: 'i64', col: 'target_project_id', cat: 'fork', catLabel: 'Fork', status: 'new', notes: 'Fork target' },
+  { api: 'milestone', type: 'Option<object>', col: 'milestone_id, milestone_title', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: 'Reuse issue milestone pattern' },
+  { api: 'merge_when_pipeline_succeeds', type: 'bool', col: 'merge_when_pipeline_succeeds', cat: 'merge', catLabel: 'Merge', status: 'new', notes: '0/1, auto-merge flag' },
+  { api: 'merge_commit_sha', type: 'Option<String>', col: 'merge_commit_sha', cat: 'merge', catLabel: 'Merge', status: 'new', notes: 'Commit ref after merge' },
+  { api: 'squash_commit_sha', type: 'Option<String>', col: 'squash_commit_sha', cat: 'merge', catLabel: 'Merge', status: 'new', notes: 'Commit ref after squash' },
+  { api: 'discussion_locked', type: 'bool', col: 'discussion_locked', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: '0/1' },
+  { api: 'should_remove_source_branch', type: 'Option<bool>', col: 'should_remove_source_branch', cat: 'branch', catLabel: 'Branch', status: 'new', notes: '0/1' },
+  { api: 'force_remove_source_branch', type: 'Option<bool>', col: 'force_remove_source_branch', cat: 'branch', catLabel: 'Branch', status: 'new', notes: '0/1' },
+  { api: 'squash', type: 'bool', col: 'squash', cat: 'merge', catLabel: 'Merge', status: 'new', notes: '0/1' },
+  { api: 'squash_on_merge', type: 'bool', col: 'squash_on_merge', cat: 'merge', catLabel: 'Merge', status: 'new', notes: '0/1' },
+  { api: 'has_conflicts', type: 'bool', col: 'has_conflicts', cat: 'merge', catLabel: 'Merge', status: 'new', notes: '0/1' },
+  { api: 'blocking_discussions_resolved', type: 'bool', col: 'blocking_discussions_resolved', cat: 'merge', catLabel: 'Merge', status: 'new', notes: '0/1' },
+  { api: 'time_stats.time_estimate', type: 'i64', col: 'time_estimate', cat: 'time', catLabel: 'Time', status: 'new', notes: 'Seconds' },
+  { api: 'time_stats.total_time_spent', type: 'i64', col: 'time_spent', cat: 'time', catLabel: 'Time', status: 'new', notes: 'Seconds' },
+  { api: 'time_stats.human_time_estimate', type: 'Option<String>', col: 'human_time_estimate', cat: 'time', catLabel: 'Time', status: 'new', notes: '' },
+  { api: 'time_stats.human_total_time_spent', type: 'Option<String>', col: 'human_time_spent', cat: 'time', catLabel: 'Time', status: 'new', notes: '' },
+  { api: 'task_completion_status.count', type: 'i64', col: 'task_count', cat: 'task', catLabel: 'Tasks', status: 'new', notes: '' },
+  { api: 'task_completion_status.completed_count', type: 'i64', col: 'task_completed_count', cat: 'task', catLabel: 'Tasks', status: 'new', notes: '' },
+  { api: 'closed_by', type: 'Option<object>', col: 'closed_by_username', cat: 'tracking', catLabel: 'Tracking', status: 'new', notes: '' },
+  { api: 'prepared_at', type: 'Option<String>', col: 'prepared_at', cat: 'schedule', catLabel: 'Schedule', status: 'new', notes: 'ms epoch, nullable' },
+  { api: 'merge_after', type: 'Option<String>', col: 'merge_after', cat: 'schedule', catLabel: 'Schedule', status: 'new', notes: 'ms epoch, nullable (scheduled merge)' },
+  { api: 'imported', type: 'bool', col: 'imported', cat: 'import', catLabel: 'Import', status: 'new', notes: '0/1' },
+  { api: 'imported_from', type: 'Option<String>', col: 'imported_from', cat: 'import', catLabel: 'Import', status: 'new', notes: '' },
+  { api: 'approvals_before_merge', type: 'Option<i64>', col: 'approvals_before_merge', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Deprecated but still returned' },
+  { api: 'references.relative', type: 'String', col: 'references_relative', cat: 'ref', catLabel: 'References', status: 'new', notes: 'Currently only short + full stored' },
+  { api: 'confidential', type: 'bool', col: 'confidential', cat: 'meta', catLabel: 'Metadata', status: 'new', notes: '0/1 (MRs can be confidential too)' },
+  { api: 'iteration.id', type: 'Option<i64>', col: 'iteration_id', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'Premium/Ultimate, null on Free' },
+  { api: 'iteration.iid', type: 'Option<i64>', col: 'iteration_iid', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'iteration.title', type: 'Option<String>', col: 'iteration_title', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'iteration.state', type: 'Option<i64>', col: 'iteration_state', cat: 'premium', catLabel: 'Premium', status: 'new', notes: '' },
+  { api: 'iteration.start_date', type: 'Option<String>', col: 'iteration_start_date', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'ISO date' },
+  { api: 'iteration.due_date', type: 'Option<String>', col: 'iteration_due_date', cat: 'premium', catLabel: 'Premium', status: 'new', notes: 'ISO date' },
+];
+
+const issueSqlLines = [
+  { type: 'comment', text: '-- Migration 007: Capture all remaining GitLab API response fields.' },
+  { type: 'comment', text: '-- Principle: mirror everything GitLab returns. No field left behind.' },
+  { type: 'blank' },
+  { type: 'divider', text: '-- ============================================================' },
+  { type: 'divider', text: '-- ISSUES: new columns' },
+  { type: 'divider', text: '-- ============================================================' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Fields currently deserialized but not stored' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN closed_at INTEGER;', note: 'ms epoch' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN author_id INTEGER;', note: 'GitLab user ID' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN author_name TEXT;', note: 'Display name' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Issue metadata' },
+  { type: 'sql', text: "ALTER TABLE issues ADD COLUMN issue_type TEXT;", note: "'issue' | 'incident' | 'test_case'" },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN confidential INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN discussion_locked INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Engagement' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN upvotes INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN downvotes INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN user_notes_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN merge_requests_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Time tracking' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN time_estimate INTEGER NOT NULL DEFAULT 0;', note: 'seconds' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN time_spent INTEGER NOT NULL DEFAULT 0;', note: 'seconds' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN human_time_estimate TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN human_time_spent TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Task lists' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN task_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN task_completed_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN has_tasks INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- References' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN references_short TEXT;', note: 'e.g. "#42"' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN references_relative TEXT;', note: 'context-dependent' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN references_full TEXT;', note: 'e.g. "group/project#42"' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Close/move tracking' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN closed_by_username TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Premium/Ultimate fields (nullable, null on Free tier)' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN weight INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN severity TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN health_status TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Import tracking' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN imported INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN imported_from TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN moved_to_id INTEGER;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Epic (Premium/Ultimate, null on Free)' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN epic_id INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN epic_iid INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN epic_title TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN epic_url TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN epic_group_id INTEGER;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Iteration (Premium/Ultimate, null on Free)' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN iteration_id INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN iteration_iid INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN iteration_title TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN iteration_state INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN iteration_start_date TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE issues ADD COLUMN iteration_due_date TEXT;' },
+];
+
+const mrSqlLines = [
+  { type: 'divider', text: '-- ============================================================' },
+  { type: 'divider', text: '-- MERGE REQUESTS: new columns' },
+  { type: 'divider', text: '-- ============================================================' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Author enrichment' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN author_id INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN author_name TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Engagement' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN upvotes INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN downvotes INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN user_notes_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Fork tracking' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN source_project_id INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN target_project_id INTEGER;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Milestone (parity with issues)' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN milestone_id INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN milestone_title TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Merge behavior' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN merge_when_pipeline_succeeds INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN merge_commit_sha TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN squash_commit_sha TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN squash INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN squash_on_merge INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Merge readiness' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN has_conflicts INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN blocking_discussions_resolved INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Branch cleanup' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN should_remove_source_branch INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN force_remove_source_branch INTEGER;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Discussion lock' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN discussion_locked INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Time tracking' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN time_estimate INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN time_spent INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN human_time_estimate TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN human_time_spent TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Task lists' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN task_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN task_completed_count INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Close tracking' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN closed_by_username TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Scheduling' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN prepared_at INTEGER;', note: 'ms epoch' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN merge_after INTEGER;', note: 'ms epoch' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- References (add relative, short + full already exist)' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN references_relative TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Import tracking' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN imported INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN imported_from TEXT;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Premium/Ultimate' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN approvals_before_merge INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN confidential INTEGER NOT NULL DEFAULT 0;' },
+  { type: 'blank' },
+  { type: 'comment', text: '-- Iteration (Premium/Ultimate, null on Free)' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN iteration_id INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN iteration_iid INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN iteration_title TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN iteration_state INTEGER;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN iteration_start_date TEXT;' },
+  { type: 'sql', text: 'ALTER TABLE merge_requests ADD COLUMN iteration_due_date TEXT;' },
+];
+
+const issueTransformRules = [
+  { from: 'closed_at', to: 'iso_to_ms() conversion (already in serde struct, just not passed through)' },
+  { from: 'time_stats', to: 'Flatten to 4 individual fields: time_estimate, time_spent, human_time_estimate, human_time_spent' },
+  { from: 'task_completion_status', to: 'Flatten to 2 individual fields: task_count, task_completed_count' },
+  { from: 'references', to: 'Flatten to 3 individual fields: references_short, references_relative, references_full' },
+  { from: 'closed_by', to: 'Extract username only (consistent with author pattern)' },
+  { from: 'author', to: 'Additionally extract id and name (currently only username)' },
+  { from: 'type / issue_type', to: 'Prefer issue_type, fall back to type field' },
+  { from: 'epic', to: 'Flatten to 5 fields: epic_id, epic_iid, epic_title, epic_url, epic_group_id' },
+  { from: 'iteration', to: 'Flatten to 6 fields: iteration_id, iteration_iid, iteration_title, iteration_state, iteration_start_date, iteration_due_date' },
+];
+
+const mrTransformRules = [
+  { from: 'milestone', to: 'Reuse upsert_milestone_tx from issue pipeline, add milestone_id + milestone_title' },
+  { from: 'prepared_at, merge_after', to: 'iso_to_ms() conversion' },
+  { from: 'source_project_id, target_project_id', to: 'Direct pass-through' },
+  { from: 'iteration', to: 'Flatten to 6 fields (same as issues)' },
+];
+
+const filesTouched = [
+  { path: 'migrations/007_complete_field_capture.sql', change: 'New file', isNew: true },
+  { path: 'src/gitlab/types.rs', change: 'Add fields to GitLabIssue, GitLabMergeRequest; add GitLabTimeStats, GitLabTaskCompletionStatus, GitLabEpic, GitLabIteration', isNew: false },
+  { path: 'src/gitlab/transformers/issue.rs', change: 'Extend IssueRow, IssueWithMetadata, transform_issue()', isNew: false },
+  { path: 'src/gitlab/transformers/merge_request.rs', change: 'Extend NormalizedMergeRequest, MergeRequestWithMetadata, transform_merge_request()', isNew: false },
+  { path: 'src/ingestion/issues.rs', change: 'Extend INSERT/UPSERT SQL, add milestone for completeness', isNew: false },
+  { path: 'src/ingestion/merge_requests.rs', change: 'Extend INSERT/UPSERT SQL, add milestone upsert', isNew: false },
+];
+
+const helperTypes = [
+  { name: 'GitLabTimeStats', fields: 'time_estimate, total_time_spent, human_time_estimate, human_total_time_spent' },
+  { name: 'GitLabTaskCompletionStatus', fields: 'count, completed_count' },
+  { name: 'GitLabReferences', fields: 'short, relative, full  (already exists for MRs, reuse for issues)' },
+  { name: 'GitLabClosedBy', fields: 'id, username, name  (reuse GitLabAuthor shape)' },
+  { name: 'GitLabEpic', fields: 'id, iid, title, url, group_id' },
+  { name: 'GitLabIteration', fields: 'id, iid, title, state, start_date, due_date' },
+];
+
+const issueStructFields = [
+  { field: 'r#type', type: 'Option<String>', comment: '#[serde(rename = "type")]' },
+  { field: 'upvotes', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'downvotes', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'user_notes_count', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'merge_requests_count', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'confidential', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'discussion_locked', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'weight', type: 'Option<i64>', comment: '' },
+  { field: 'time_stats', type: 'Option<GitLabTimeStats>', comment: '' },
+  { field: 'task_completion_status', type: 'Option<GitLabTaskCompletionStatus>', comment: '' },
+  { field: 'has_tasks', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'references', type: 'Option<GitLabReferences>', comment: '' },
+  { field: 'closed_by', type: 'Option<GitLabAuthor>', comment: '' },
+  { field: 'severity', type: 'Option<String>', comment: '' },
+  { field: 'health_status', type: 'Option<String>', comment: '' },
+  { field: 'imported', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'imported_from', type: 'Option<String>', comment: '' },
+  { field: 'moved_to_id', type: 'Option<i64>', comment: '' },
+  { field: 'issue_type', type: 'Option<String>', comment: 'alias for type' },
+  { field: 'epic', type: 'Option<GitLabEpic>', comment: '' },
+  { field: 'iteration', type: 'Option<GitLabIteration>', comment: '' },
+];
+
+const mrStructFields = [
+  { field: 'upvotes', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'downvotes', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'user_notes_count', type: 'i64', comment: '#[serde(default)]' },
+  { field: 'source_project_id', type: 'Option<i64>', comment: '' },
+  { field: 'target_project_id', type: 'Option<i64>', comment: '' },
+  { field: 'milestone', type: 'Option<GitLabMilestone>', comment: 'reuse existing type' },
+  { field: 'merge_when_pipeline_succeeds', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'merge_commit_sha', type: 'Option<String>', comment: '' },
+  { field: 'squash_commit_sha', type: 'Option<String>', comment: '' },
+  { field: 'squash', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'squash_on_merge', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'has_conflicts', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'blocking_discussions_resolved', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'should_remove_source_branch', type: 'Option<bool>', comment: '' },
+  { field: 'force_remove_source_branch', type: 'Option<bool>', comment: '' },
+  { field: 'discussion_locked', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'time_stats', type: 'Option<GitLabTimeStats>', comment: '' },
+  { field: 'task_completion_status', type: 'Option<GitLabTaskCompletionStatus>', comment: '' },
+  { field: 'closed_by', type: 'Option<GitLabAuthor>', comment: '' },
+  { field: 'prepared_at', type: 'Option<String>', comment: '' },
+  { field: 'merge_after', type: 'Option<String>', comment: '' },
+  { field: 'imported', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'imported_from', type: 'Option<String>', comment: '' },
+  { field: 'approvals_before_merge', type: 'Option<i64>', comment: '' },
+  { field: 'confidential', type: 'bool', comment: '#[serde(default)]' },
+  { field: 'iteration', type: 'Option<GitLabIteration>', comment: '' },
+];
+
+// ============================================================
+// RENDER
+// ============================================================
+
+function esc(s) {
+  return s.replace(/&/g,'&amp;').replace(/</g,'&lt;').replace(/>/g,'&gt;');
+}
+
+function renderFieldTable(fields, tbodyId) {
+  const tbody = document.querySelector(`#${tbodyId} tbody`);
+  fields.forEach(f => {
+    const tr = document.createElement('tr');
+    tr.dataset.status = f.status;
+    tr.dataset.search = `${f.api} ${f.col} ${f.cat} ${f.notes}`.toLowerCase();
+    const dotClass = f.status === 'wired' ? 'dot-yellow' : 'dot-green';
+    const statusText = f.status === 'wired' ? 'Wire up' : 'New capture';
+    tr.innerHTML = `
+      <td><code>${esc(f.api)}</code></td>
+      <td style="color:var(--text-dim); font-size:12px; font-family:monospace;">${esc(f.type)}</td>
+      <td><code>${esc(f.col)}</code></td>
+      <td><span class="cat cat-${f.cat}">${esc(f.catLabel)}</span></td>
+      <td><span class="dot ${dotClass}"></span>${statusText}</td>
+      <td style="color:var(--text-dim); font-size:12px;">${esc(f.notes)}</td>
+    `;
+    tbody.appendChild(tr);
+  });
+}
+
+function renderSqlBlock(lines, containerId) {
+  const container = document.getElementById(containerId);
+  let count = 0;
+  const html = lines.map(l => {
+    if (l.type === 'blank') return '<span class="sql-line">&nbsp;</span>';
+    if (l.type === 'comment') return `<span class="sql-line"><span class="sql-comment">${esc(l.text)}</span></span>`;
+    if (l.type === 'divider') return `<span class="sql-line"><span class="sql-divider">${esc(l.text)}</span></span>`;
+    count++;
+    let formatted = l.text
+      .replace(/^(ALTER TABLE)/,  '<span class="sql-keyword">$1</span>')
+      .replace(/(ADD COLUMN)/, '<span class="sql-keyword">$1</span>')
+      .replace(/(INTEGER|TEXT)/, '<span class="sql-type">$1</span>')
+      .replace(/(NOT NULL DEFAULT \d+)/, '<span class="sql-default">$1</span>')
+      .replace(/(NOT NULL DEFAULT 0)/, '<span class="sql-default">$1</span>');
+    const note = l.note ? `  <span class="sql-comment">-- ${esc(l.note)}</span>` : '';
+    return `<span class="sql-line">${formatted}${note}</span>`;
+  }).join('\n');
+  container.innerHTML = html;
+  return count;
+}
+
+function renderStructBlock(fields, containerId) {
+  const container = document.getElementById(containerId);
+  const html = fields.map(f => {
+    const comment = f.comment ? `  <span class="struct-comment">// ${esc(f.comment)}</span>` : '';
+    return `<span class="struct-line"><span class="struct-field">${esc(f.field)}</span>: <span class="struct-type">${esc(f.type)}</span>,${comment}</span>`;
+  }).join('\n');
+  container.innerHTML = html;
+}
+
+function renderHelperTypes() {
+  const container = document.getElementById('helperTypes');
+  const html = helperTypes.map(t => {
+    return `<span class="struct-line"><span class="struct-type">${esc(t.name)}</span> <span class="struct-comment">{ ${esc(t.fields)} }</span></span>`;
+  }).join('\n');
+  container.innerHTML = html;
+}
+
+function renderTransforms(rules, containerId) {
+  const container = document.getElementById(containerId);
+  rules.forEach(r => {
+    const div = document.createElement('div');
+    div.className = 'transform-rule';
+    div.innerHTML = `
+      <span class="transform-from"><code>${esc(r.from)}</code></span>
+      <span class="transform-arrow">&rarr;</span>
+      <span class="transform-to">${esc(r.to)}</span>
+    `;
+    container.appendChild(div);
+  });
+}
+
+function renderFiles() {
+  const container = document.getElementById('filesList');
+  filesTouched.forEach(f => {
+    const div = document.createElement('div');
+    div.className = 'file-item';
+    const iconClass = f.isNew ? 'file-icon-new' : 'file-icon-mod';
+    const icon = f.isNew ? '+' : '~';
+    div.innerHTML = `
+      <div class="file-icon ${iconClass}">${icon}</div>
+      <div>
+        <div class="file-path">${esc(f.path)}</div>
+        <div class="file-desc">${esc(f.change)}</div>
+      </div>
+    `;
+    container.appendChild(div);
+  });
+}
+
+// ============================================================
+// TABS
+// ============================================================
+
+document.querySelectorAll('.tab').forEach(tab => {
+  tab.addEventListener('click', () => {
+    document.querySelectorAll('.tab').forEach(t => t.classList.remove('active'));
+    document.querySelectorAll('.tab-content').forEach(tc => tc.classList.remove('active'));
+    tab.classList.add('active');
+    document.getElementById('tab-' + tab.dataset.tab).classList.add('active');
+  });
+});
+
+// ============================================================
+// FILTERS
+// ============================================================
+
+function setupFilters(tableId, searchId, filterTable) {
+  const search = document.getElementById(searchId);
+  const rows = document.querySelectorAll(`#${tableId} tbody tr`);
+
+  function applyFilters() {
+    const query = search.value.toLowerCase();
+    const activeFilter = document.querySelector(`.filter-btn.active[data-table="${filterTable}"]`);
+    const filter = activeFilter ? activeFilter.dataset.filter : 'all';
+    rows.forEach(row => {
+      const matchesSearch = !query || row.dataset.search.includes(query);
+      const matchesFilter = filter === 'all' || row.dataset.status === filter;
+      row.style.display = (matchesSearch && matchesFilter) ? '' : 'none';
+    });
+  }
+
+  search.addEventListener('input', applyFilters);
+  document.querySelectorAll(`.filter-btn[data-table="${filterTable}"]`).forEach(btn => {
+    btn.addEventListener('click', () => {
+      document.querySelectorAll(`.filter-btn[data-table="${filterTable}"]`).forEach(b => b.classList.remove('active'));
+      btn.classList.add('active');
+      applyFilters();
+    });
+  });
+}
+
+// ============================================================
+// INIT
+// ============================================================
+
+renderFieldTable(issueFields, 'issueTable');
+renderFieldTable(mrFields, 'mrTable');
+const issueSqlCount = renderSqlBlock(issueSqlLines, 'issueSql');
+const mrSqlCount = renderSqlBlock(mrSqlLines, 'mrSql');
+renderStructBlock(issueStructFields, 'issueStruct');
+renderStructBlock(mrStructFields, 'mrStruct');
+renderHelperTypes();
+renderTransforms(issueTransformRules, 'issueTransforms');
+renderTransforms(mrTransformRules, 'mrTransforms');
+renderFiles();
+
+setupFilters('issueTable', 'issueSearch', 'issues');
+setupFilters('mrTable', 'mrSearch', 'mrs');
+
+// Counts
+document.getElementById('issueCount').textContent = issueFields.length;
+document.getElementById('mrCount').textContent = mrFields.length;
+document.getElementById('issueSqlCount').textContent = issueSqlCount + ' ALTER statements';
+document.getElementById('mrSqlCount').textContent = mrSqlCount + ' ALTER statements';
+document.getElementById('issueFieldCount').textContent = issueStructFields.length + ' fields';
+document.getElementById('mrFieldCount').textContent = mrStructFields.length + ' fields';
+
+document.getElementById('issueNewCols').textContent = issueSqlCount;
+document.getElementById('mrNewCols').textContent = mrSqlCount;
+document.getElementById('totalNewCols').textContent = issueSqlCount + mrSqlCount;
+document.getElementById('issueNewNum').textContent = issueSqlCount;
+document.getElementById('mrNewNum').textContent = mrSqlCount;
+
+// Bars
+const issueExisting = 18;
+const mrExisting = 29;
+const issueTotal = issueExisting + issueSqlCount;
+const mrTotal = mrExisting + mrSqlCount;
+document.getElementById('issueBarExisting').style.width = (issueExisting / issueTotal * 100) + '%';
+document.getElementById('issueBarNew').style.width = (issueSqlCount / issueTotal * 100) + '%';
+document.getElementById('mrBarExisting').style.width = (mrExisting / mrTotal * 100) + '%';
+document.getElementById('mrBarNew').style.width = (mrSqlCount / mrTotal * 100) + '%';
+</script>
+</body>
+</html>
diff --git a/skills/agent-swarm-launcher/SKILL.md b/skills/agent-swarm-launcher/SKILL.md
new file mode 100644
index 0000000..c01c77a
--- /dev/null
+++ b/skills/agent-swarm-launcher/SKILL.md
@@ -0,0 +1,36 @@
+---
+name: agent-swarm-launcher
+description: "Launch a multi-agent “swarm” workflow for a repository: read and follow AGENTS.md/README.md, perform an architecture/codebase reconnaissance, then coordinate work via Agent Mail / beads-style task tracking when those tools are available. Use when you want to quickly bootstrap a coordinated agent workflow, avoid communication deadlocks, and start making progress on prioritized tasks."
+---
+
+# Agent Swarm Launcher
+
+## Workflow (do in order)
+
+1. Read *all* `AGENTS.md` and `README.md` files carefully and completely.
+   - If multiple `AGENTS.md` files exist, treat deeper ones as higher priority within their directory scope.
+   - Note any required workflows (e.g., TDD), tooling conventions, and “robot mode” flags.
+
+2. Enter “code investigation” mode and understand the project.
+   - Identify entrypoints, key packages/modules, and how data flows.
+   - Note build/test commands and any local dev constraints.
+   - Summarize the technical architecture and purpose of the project.
+
+3. Register with Agent Mail and coordinate, if available.
+   - If “MCP Agent Mail” exists in this environment, register and introduce yourself to the other agents.
+   - Check Agent Mail and promptly respond to any messages.
+   - If “beads” tracking is used by the repo/team, open/continue the current bead(s) and mark progress as you go.
+   - If Agent Mail/beads are not available, state that plainly and proceed with a lightweight local substitute (a short task checklist in the thread).
+
+4. Start work (do not get stuck waiting).
+   - Acknowledge incoming requests promptly.
+   - Do not get stuck in “communication purgatory” where nothing is getting done.
+   - If you are blocked on prioritization, look for a prioritization tool mentioned in `AGENTS.md` (for example “bv”) and use it; otherwise propose the next best task(s) and proceed.
+   - If `AGENTS.md` references a task system (e.g., beads), pick the next task you can complete usefully and start.
+
+## Execution rules
+
+- Follow repository instructions over this skill if they conflict.
+- Prefer action + short status updates over prolonged coordination.
+- If a referenced tool does not exist, do not hallucinate it—fall back and keep moving.
+- Do not claim you registered with or heard from other agents unless you actually did via the available tooling.
diff --git a/skills/agent-swarm-launcher/agents/openai.yaml b/skills/agent-swarm-launcher/agents/openai.yaml
new file mode 100644
index 0000000..0e714d4
--- /dev/null
+++ b/skills/agent-swarm-launcher/agents/openai.yaml
@@ -0,0 +1,4 @@
+interface:
+  display_name: "Agent Swarm Launcher"
+  short_description: "Kick off multi-agent repo onboarding"
+  default_prompt: "Use $agent-swarm-launcher to onboard, coordinate, and start the next prioritized task."