From 77ff8283fe189f89202c57cf9eebc4fdf11610c8 Mon Sep 17 00:00:00 2001 From: Steve Yegge Date: Tue, 23 Dec 2025 12:40:03 -0800 Subject: [PATCH] bd sync: 2025-12-23 12:40:03 --- .beads/issues.jsonl | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/.beads/issues.jsonl b/.beads/issues.jsonl index 39235b95..3debac0e 100644 --- a/.beads/issues.jsonl +++ b/.beads/issues.jsonl @@ -367,7 +367,7 @@ {"id":"bd-pdr2","title":"Consider backwards compatibility for ready() and list() return type change","description":"PR #481 changed the return types of `ready()` and `list()` from `list[Issue]` to `list[IssueMinimal] | CompactedResult`. This is a breaking change for MCP clients.\n\n## Impact Assessment\nBreaking change affects:\n- Any MCP client expecting `list[Issue]` from ready()\n- Any MCP client expecting `list[Issue]` from list()\n- Client code that accesses full Issue fields (description, design, acceptance_criteria, timestamps, dependencies, dependents)\n\n## Current Behavior\n- ready() returns `list[IssueMinimal] | CompactedResult`\n- list() returns `list[IssueMinimal] | CompactedResult`\n- show() still returns full `Issue` (good)\n\n## Considerations\n**Pros of current approach:**\n- Forces clients to use show() for full details (good for context efficiency)\n- Simple mental model (always use show for full data)\n- Documentation warns about this\n\n**Cons:**\n- Clients expecting list[Issue] will break\n- No graceful degradation option\n- No migration period\n\n## Potential Solutions\n1. Add optional parameter `full_details=false` to ready/list (would increase payload)\n2. Create separate tools: ready_minimal/list_minimal + ready_full/list_full\n3. Accept breaking change and document upgrade path (current approach)\n4. Version the MCP server and document migration guide\n\n## Recommendation\nCurrent approach (solution 3) is reasonable if:\n- Changelog clearly documents the breaking change\n- Migration guide provided to clients\n- Error handling is graceful for clients expecting specific fields","status":"open","priority":2,"issue_type":"task","created_at":"2025-12-14T14:24:56.460465-08:00","updated_at":"2025-12-14T14:24:56.460465-08:00","dependencies":[{"issue_id":"bd-pdr2","depends_on_id":"bd-otf4","type":"discovered-from","created_at":"2025-12-14T14:24:56.461959-08:00","created_by":"stevey","metadata":"{}"}]} {"id":"bd-pe4s","title":"JSON test issue","description":"Line 1\nLine 2\nLine 3","status":"tombstone","priority":2,"issue_type":"task","created_at":"2025-12-16T16:14:36.969074-08:00","updated_at":"2025-12-17T16:11:17.070763-08:00","deleted_at":"2025-12-17T16:11:17.070763-08:00","deleted_by":"batch delete","delete_reason":"batch delete","original_type":"task"} {"id":"bd-pgcs","title":"Clean up orphaned child issues (bd-cb64c226.*, bd-cbed9619.*)","description":"## Problem\n\nEvery bd command shows warnings about 12 orphaned child issues:\n- bd-cb64c226.1, .6, .8, .9, .10, .12, .13\n- bd-cbed9619.1, .2, .3, .4, .5\n\nThese are hierarchical IDs (parent.child format) where the parent issues no longer exist.\n\n## Impact\n\n- Clutters output of every bd command\n- Confusing for users\n- Indicates incomplete cleanup of deleted parent issues\n\n## Proposed Solution\n\n1. Delete the orphaned issues since their parents no longer exist:\n ```bash\n bd delete bd-cb64c226.1 bd-cb64c226.6 bd-cb64c226.8 ...\n ```\n\n2. Or convert them to top-level issues if they contain useful content\n\n## Investigation Needed\n\n- What were the parent issues bd-cb64c226 and bd-cbed9619?\n- Why were they deleted without their children?\n- Should bd delete cascade to children automatically?","status":"tombstone","priority":2,"issue_type":"task","created_at":"2025-12-16T23:06:17.240571-08:00","updated_at":"2025-12-17T16:11:17.070763-08:00","deleted_at":"2025-12-17T16:11:17.070763-08:00","deleted_by":"batch delete","delete_reason":"batch delete","original_type":"task"} -{"id":"bd-phtv","title":"bd pin: pinned field overwritten by subsequent bd commands","description":"## Summary\n\nThe `bd pin` command correctly sets `pinned=1` in SQLite, but any subsequent `bd` command (including read-only commands like `bd show`) resets `pinned` to 0. This happens even with `--no-auto-import`, `--no-auto-flush`, and `--sandbox` flags.\n\n## Reproduction Steps\n\n```bash\n# 1. Pin an issue\n$ bd --no-daemon pin \u003cissue-id\u003e --for=max\nšŸ“Œ Pinned \u003cissue-id\u003e to max's hook\n\n# 2. Verify SQLite has pinned=1\n$ sqlite3 .beads/beads.db 'SELECT id, pinned FROM issues WHERE id=\"\u003cissue-id\u003e\"'\n\u003cissue-id\u003e|1 # āœ“ Correct\n\n# 3. Run any bd command (even read-only)\n$ bd --no-daemon --no-auto-import show \u003cissue-id\u003e --json\n[output shows issue without pinned field]\n\n# 4. Check SQLite again\n$ sqlite3 .beads/beads.db 'SELECT id, pinned FROM issues WHERE id=\"\u003cissue-id\u003e\"'\n\u003cissue-id\u003e|0 # āœ— WRONG - pinned was reset!\n```\n\n## Expected Behavior\n\nThe `pinned` field should persist in SQLite across bd command invocations.\n\n## Actual Behavior\n\nThe `pinned` field is reset to 0 (or NULL) by subsequent bd commands.\n\n## Investigation Findings\n\n1. **Not auto-import**: Adding `--no-auto-import` doesn't help. Debug output shows 'auto-import skipped, JSONL unchanged (hash match)' but pinned is still reset.\n\n2. **Not sandbox mode**: Adding `--sandbox` doesn't help either.\n\n3. **JSONL export issue**: The `pinned` field uses `json:\"pinned,omitempty\"` tag in types.go. When pinned=false, it's omitted from JSONL. But this shouldn't cause the overwrite since we're not auto-importing.\n\n4. **Something else is importing**: There must be another code path that reads from JSONL and overwrites SQLite, bypassing the --no-auto-import flag.\n\n## Possible Causes\n\n1. **HydrateFromMultiRepo**: Called in daemon.go and daemon_sync.go - might be running during store initialization\n2. **Config redirect handling**: The .beads/redirect mechanism might cause double-loading\n3. **WAL mode issues**: SQLite WAL might not be flushing between commands\n4. **Store initialization**: Something in ensureStoreActive() or NewSQLiteStorage() might be importing\n\n## Impact\n\n- `bd hook` command can't show pinned work (cosmetic but confusing)\n- `bd list --pinned` returns nothing\n- Gas Town gt sling command works around this by using AttachMolecule() mechanism instead\n\n## Workaround\n\nThe handoff bead attachment mechanism (AttachMolecule storing molecule ID in description field) works correctly. This is the primary work assignment mechanism in Gas Town.\n\n## Related\n\n- gt-o3is: Gas Town issue that discovered this bug\n- bd-7h5: Original pinned column implementation","status":"open","priority":1,"issue_type":"bug","created_at":"2025-12-23T12:32:20.046988-08:00","updated_at":"2025-12-23T12:32:20.046988-08:00","labels":["export:pinned-field-fix"]} +{"id":"bd-phtv","title":"bd pin: pinned field overwritten by subsequent bd commands","description":"## Summary\n\nThe `bd pin` command correctly sets `pinned=1` in SQLite, but any subsequent `bd` command (including read-only commands like `bd show`) resets `pinned` to 0.\n\n## Reproduction Steps\n\n```bash\nbd --no-daemon pin \u003cissue-id\u003e --for=max\nsqlite3 .beads/beads.db \"SELECT id, pinned FROM issues WHERE id=\\\"\u003cissue-id\u003e\\\"\"\n# Shows pinned=1 āœ“\n\nbd --no-daemon show \u003cissue-id\u003e --json\nsqlite3 .beads/beads.db \"SELECT id, pinned FROM issues WHERE id=\\\"\u003cissue-id\u003e\\\"\"\n# Shows pinned=0 āœ— WRONG\n```\n\n## Root Cause Investigation\n\n### Prime Suspects\n\n1. **JSONL import overwrites DB** - The `pinned` field has `omitempty` so false values arent in JSONL. When JSONL is imported, it overwrites the DB pinned=1 with default pinned=0.\n\n2. **Files to check:**\n - `internal/importer/importer.go` - ImportIssue() may unconditionally set all fields\n - `internal/storage/sqlite/issues.go` - UpsertIssue() may not preserve pinned\n - `cmd/bd/main.go` - ensureStoreActive() may trigger import\n\n### Debug Steps\n\n```bash\n# Add debug logging to track what is writing pinned=0\ngrep -rn \"pinned\" internal/storage/sqlite/*.go\ngrep -rn \"Pinned\" internal/importer/*.go\n```\n\n## Likely Fix\n\nIn `internal/importer/importer.go` or `internal/storage/sqlite/issues.go`:\n\n```go\n// When upserting from JSONL, preserve pinned field if already set\nfunc (s *SQLiteStorage) UpsertIssue(ctx context.Context, issue *types.Issue) error {\n // Check if issue exists and is pinned\n existing, _ := s.GetIssue(ctx, issue.ID)\n if existing != nil \u0026\u0026 existing.Pinned \u0026\u0026 !issue.Pinned {\n // Preserve existing pinned status\n issue.Pinned = existing.Pinned\n }\n // ... rest of upsert\n}\n```\n\nOR the import should skip fields that are omitempty and not present in JSONL:\n\n```go\n// In importer, only update fields that are explicitly set in JSONL\n// Pinned with omitempty means absent = dont change, not absent = false\n```\n\n## Testing\n\n```bash\n# After fix:\nbd --no-daemon pin \u003cissue-id\u003e --for=max\nbd --no-daemon show \u003cissue-id\u003e --json # Should not reset pinned\nbd list --pinned # Should show the pinned issue\nbd hook --agent max # Should show pinned work\n```\n\n## Files to Modify\n\n1. **internal/importer/importer.go** - Preserve pinned on import\n2. **internal/storage/sqlite/issues.go** - UpsertIssue preserve pinned\n3. **Add test** in internal/importer/importer_test.go\n\n## Success Criteria\n- `bd pin` survives subsequent bd commands\n- `bd list --pinned` shows pinned issues\n- `bd hook --agent X` shows pinned work\n- Existing tests still pass","status":"open","priority":1,"issue_type":"bug","created_at":"2025-12-23T12:32:20.046988-08:00","updated_at":"2025-12-23T12:39:47.371978-08:00","labels":["export:pinned-field-fix"]} {"id":"bd-phwd","title":"Add timeout message for long-running git push operations","description":"When git push hangs waiting for credential/browser auth, show a periodic message to the user instead of appearing frozen. Add timeout messaging after N seconds of inactivity during git operations.","status":"closed","priority":2,"issue_type":"task","created_at":"2025-12-21T11:44:57.318984535-07:00","updated_at":"2025-12-21T11:46:05.218023559-07:00","closed_at":"2025-12-21T11:46:05.218023559-07:00"} {"id":"bd-pn0t","title":"Add 0.33.2 to versionChanges in info.go","description":"Add new entry at TOP of versionChanges in cmd/bd/info.go with release notes from CHANGELOG.md. Must do before bump-version.sh --commit.","status":"tombstone","priority":1,"issue_type":"task","created_at":"2025-12-21T16:10:13.760056-08:00","updated_at":"2025-12-21T17:29:31.791368-08:00","deleted_at":"2025-12-21T17:29:31.791368-08:00","deleted_by":"batch delete","delete_reason":"batch delete","original_type":"task","wisp":true} {"id":"bd-psg","title":"Add tests for dependency management","description":"Key dependency functions like mergeBidirectionalTrees, GetDependencyTree, and DetectCycles have low or no coverage. These are essential for maintaining data integrity in the dependency graph.","status":"closed","priority":1,"issue_type":"task","created_at":"2025-12-18T07:00:43.458548462-07:00","updated_at":"2025-12-19T09:54:57.018745301-07:00","closed_at":"2025-12-18T10:24:56.271508339-07:00","dependencies":[{"issue_id":"bd-psg","depends_on_id":"bd-6ss","type":"discovered-from","created_at":"2025-12-18T07:00:43.463910911-07:00","created_by":"matt"}]} @@ -436,6 +436,7 @@ {"id":"bd-u2sc.3","title":"Split large cmd/bd files into logical modules","description":"Split large cmd/bd files into logical modules for maintainability.\n\n## Current State\n| File | Lines | Status |\n|------|-------|--------|\n| sync.go | 2203 | Too large |\n| init.go | 1742 | Too large |\n| show.go | 1419 | Too large |\n| compact.go | 1190 | Borderline |\n\n## Proposed Splits\n\n### 1. sync.go (2203 lines) ā†’ 4 files\n\n**sync.go** (~500 lines) - Main command and coordination\n```go\nvar syncCmd = \u0026cobra.Command{...}\nfunc init() {...}\nfunc doSync(...) {...}\n```\n\n**sync_export.go** (~400 lines) - Export operations\n```go\nfunc exportToJSONL(...)\nfunc markDirtyAndScheduleFlush(...)\nfunc flushPendingChanges(...)\n```\n\n**sync_import.go** (~500 lines) - Import operations\n```go\nfunc importFromJSONL(...)\nfunc handleImportConflicts(...)\nfunc mergeImportedIssues(...)\n```\n\n**sync_branch.go** (~400 lines) - Git branch operations\n```go\nfunc commitToSyncBranch(...)\nfunc pullFromSyncBranch(...)\nfunc handleSyncBranchDivergence(...)\n```\n\n### 2. init.go (1742 lines) ā†’ 3 files\n\n**init.go** (~400 lines) - Main init command\n```go\nvar initCmd = \u0026cobra.Command{...}\nfunc runInit(...)\nfunc determinePrefix(...)\n```\n\n**init_wizard.go** (~500 lines) - Interactive setup\n```go\nfunc runContributorWizard(...)\nfunc runTeamWizard(...)\nfunc promptForConfig(...)\n```\n\n**init_hooks.go** (~400 lines) - Git hooks setup\n```go\nfunc installGitHooks(...)\nfunc configureAutosync(...)\nfunc setupMergeDriver(...)\n```\n\n### 3. show.go (1419 lines) ā†’ 3 files\n\n**show.go** (~400 lines) - Main show command\n```go\nvar showCmd = \u0026cobra.Command{...}\nfunc showIssue(...)\nfunc showMultipleIssues(...)\n```\n\n**show_format.go** (~400 lines) - Output formatting\n```go\nfunc formatIssueText(...)\nfunc formatIssueMarkdown(...)\nfunc formatDependencyTree(...)\n```\n\n**show_threads.go** (~300 lines) - Thread/conversation display\n```go\nfunc showThread(...)\nfunc formatThreadMessages(...)\n```\n\n### 4. compact.go (1190 lines) ā†’ 2 files\n\n**compact.go** (~600 lines) - Main compact command\n```go\nvar compactCmd = \u0026cobra.Command{...}\nfunc runCompact(...)\n```\n\n**compact_tiers.go** (~400 lines) - Tier-specific logic\n```go\nfunc compactTier1(...)\nfunc compactTier2(...)\nfunc squashWisps(...)\n```\n\n## Implementation Steps\n\n1. **Start with sync.go** (largest file)\n2. **Create new files** with same package declaration\n3. **Move functions** maintaining related code together\n4. **Update any file-local variables** that need to be shared\n5. **Run tests** after each split:\n ```bash\n go test -short ./cmd/bd/...\n ```\n\n## File Organization After Split\n```\ncmd/bd/\nā”œā”€ā”€ sync.go (~500 lines)\nā”œā”€ā”€ sync_export.go (~400 lines)\nā”œā”€ā”€ sync_import.go (~500 lines)\nā”œā”€ā”€ sync_branch.go (~400 lines)\nā”œā”€ā”€ init.go (~400 lines)\nā”œā”€ā”€ init_wizard.go (~500 lines)\nā”œā”€ā”€ init_hooks.go (~400 lines)\nā”œā”€ā”€ show.go (~400 lines)\nā”œā”€ā”€ show_format.go (~400 lines)\nā”œā”€ā”€ show_threads.go (~300 lines)\nā”œā”€ā”€ compact.go (~600 lines)\nā””ā”€ā”€ compact_tiers.go (~400 lines)\n```\n\n## Success Criteria\n- No file \u003e 600 lines\n- All tests pass\n- Related code stays together\n- Clear file naming indicates purpose","status":"open","priority":3,"issue_type":"task","created_at":"2025-12-22T14:27:06.146343-08:00","updated_at":"2025-12-23T12:33:22.187826-08:00","dependencies":[{"issue_id":"bd-u2sc.3","depends_on_id":"bd-u2sc","type":"parent-child","created_at":"2025-12-22T14:27:06.146704-08:00","created_by":"daemon"}]} {"id":"bd-u2sc.4","title":"Introduce slog for structured daemon logging","description":"Introduce slog for structured daemon logging.\n\n## Current State\nDaemon uses fmt.Fprintf for logging:\n```go\nfmt.Fprintf(os.Stderr, \"Warning: failed to detect user role: %v\\n\", err)\nfmt.Fprintf(logFile, \"[%s] %s\\n\", time.Now().Format(time.RFC3339), msg)\n```\n\nThis produces unstructured, hard-to-parse logs.\n\n## Target State\nUse Go 1.21+ slog for structured logging:\n```go\nslog.Warn(\"failed to detect user role\", \"error\", err)\nslog.Info(\"sync completed\", \"created\", 5, \"updated\", 3, \"duration_ms\", 150)\n```\n\n## Implementation\n\n### 1. Create logger setup (internal/daemon/logger.go)\n\n```go\npackage daemon\n\nimport (\n \"io\"\n \"log/slog\"\n \"os\"\n)\n\n// SetupLogger configures the daemon logger.\n// Returns a cleanup function to close the log file.\nfunc SetupLogger(logPath string, jsonFormat bool, level slog.Level) (func(), error) {\n var w io.Writer = os.Stderr\n var cleanup func()\n \n if logPath \\!= \"\" {\n f, err := os.OpenFile(logPath, os.O_CREATE|os.O_APPEND|os.O_WRONLY, 0600)\n if err \\!= nil {\n return nil, err\n }\n w = io.MultiWriter(os.Stderr, f)\n cleanup = func() { f.Close() }\n }\n \n var handler slog.Handler\n opts := \u0026slog.HandlerOptions{Level: level}\n \n if jsonFormat {\n handler = slog.NewJSONHandler(w, opts)\n } else {\n handler = slog.NewTextHandler(w, opts)\n }\n \n slog.SetDefault(slog.New(handler))\n \n return cleanup, nil\n}\n```\n\n### 2. Add log level flag\n\nIn cmd/bd/daemon.go:\n```go\ndaemonCmd.Flags().String(\"log-level\", \"info\", \"Log level (debug, info, warn, error)\")\ndaemonCmd.Flags().Bool(\"log-json\", false, \"Output logs in JSON format\")\n```\n\n### 3. Replace fmt.Fprintf with slog calls\n\n**Pattern 1: Simple messages**\n```go\n// Before\nfmt.Fprintf(os.Stderr, \"Starting daemon on %s\\n\", socketPath)\n\n// After\nslog.Info(\"starting daemon\", \"socket\", socketPath)\n```\n\n**Pattern 2: Errors**\n```go\n// Before\nfmt.Fprintf(os.Stderr, \"Error: failed to connect: %v\\n\", err)\n\n// After\nslog.Error(\"failed to connect\", \"error\", err)\n```\n\n**Pattern 3: Debug info**\n```go\n// Before\nif debug {\n fmt.Fprintf(os.Stderr, \"Received request: %s\\n\", method)\n}\n\n// After\nslog.Debug(\"received request\", \"method\", method)\n```\n\n**Pattern 4: Structured data**\n```go\n// Before\nfmt.Fprintf(logFile, \"Import: %d created, %d updated\\n\", created, updated)\n\n// After\nslog.Info(\"import completed\", \n \"created\", created,\n \"updated\", updated,\n \"unchanged\", unchanged,\n \"duration_ms\", duration.Milliseconds())\n```\n\n### 4. Files to Update\n\n| File | Changes |\n|------|---------|\n| cmd/bd/daemon.go | Add log flags, call SetupLogger |\n| internal/daemon/server.go | Replace fmt with slog |\n| internal/daemon/rpc_handler.go | Replace fmt with slog |\n| internal/daemon/sync.go | Replace fmt with slog |\n| internal/daemon/autoflush.go | Replace fmt with slog |\n| internal/daemon/logger.go | New file |\n\n### 5. Log output examples\n\n**Text format (default):**\n```\ntime=2025-12-23T12:30:00Z level=INFO msg=\"daemon started\" socket=/tmp/bd.sock pid=12345\ntime=2025-12-23T12:30:01Z level=INFO msg=\"import completed\" created=5 updated=3 duration_ms=150\ntime=2025-12-23T12:30:05Z level=WARN msg=\"sync branch diverged\" local_ahead=2 remote_ahead=1\n```\n\n**JSON format (--log-json):**\n```json\n{\"time\":\"2025-12-23T12:30:00Z\",\"level\":\"INFO\",\"msg\":\"daemon started\",\"socket\":\"/tmp/bd.sock\",\"pid\":12345}\n{\"time\":\"2025-12-23T12:30:01Z\",\"level\":\"INFO\",\"msg\":\"import completed\",\"created\":5,\"updated\":3,\"duration_ms\":150}\n```\n\n## Migration Strategy\n\n1. **Add logger.go** with SetupLogger\n2. **Update daemon startup** to initialize slog\n3. **Convert one file at a time** (start with server.go)\n4. **Test after each file**\n5. **Remove old logging code** once all converted\n\n## Testing\n\n```bash\n# Start daemon with debug logging\nbd daemon start --log-level debug\n\n# Check logs\nbd daemons logs . | head -20\n\n# Test JSON output\nbd daemon start --log-json --log-level debug\nbd daemons logs . | jq .\n```\n\n## Success Criteria\n- All daemon logging uses slog\n- --log-level controls verbosity\n- --log-json produces machine-parseable output\n- Log entries have consistent structure\n- No fmt.Fprintf to stderr in daemon code","status":"open","priority":3,"issue_type":"task","created_at":"2025-12-22T14:27:16.47144-08:00","updated_at":"2025-12-23T12:36:19.089183-08:00","dependencies":[{"issue_id":"bd-u2sc.4","depends_on_id":"bd-u2sc","type":"parent-child","created_at":"2025-12-22T14:27:16.471878-08:00","created_by":"daemon"}]} {"id":"bd-u66e","title":"Implement bd gate create/show/list/close/wait commands","description":"Implement the gate CLI commands.\n\n## Commands\n```bash\n# Create gate (returns gate ID)\nbd gate create --await \u003ctype\u003e:\u003cid\u003e --timeout \u003cduration\u003e --notify \u003caddr\u003e\n\n# Check gate status\nbd gate show \u003cid\u003e\n\n# List open gates\nbd gate list\n\n# Close gate (usually done by Deacon)\nbd gate close \u003cid\u003e --reason \"completed\"\n\n# Add waiter to existing gate\nbd gate wait \u003cid\u003e --notify \u003caddr\u003e\n```\n\n## Implementation\n- Add cmd/bd/gate.go with subcommands\n- Gate create creates wisp issue of type gate\n- Gate list filters for open gates\n- Gate wait adds to waiters[] array\n- All support --json output","status":"closed","priority":1,"issue_type":"task","created_at":"2025-12-23T11:44:34.022464-08:00","updated_at":"2025-12-23T12:06:18.550673-08:00","closed_at":"2025-12-23T12:06:18.550673-08:00","close_reason":"Implemented bd gate commands: create, show, list, close, wait. All support --json output and work with wisp/gate fields.","dependencies":[{"issue_id":"bd-u66e","depends_on_id":"bd-udsi","type":"parent-child","created_at":"2025-12-23T11:44:52.823353-08:00","created_by":"daemon"},{"issue_id":"bd-u66e","depends_on_id":"bd-lz49","type":"blocks","created_at":"2025-12-23T11:44:56.349662-08:00","created_by":"daemon"}]} +{"id":"bd-ucgz","title":"Migration invariants should exclude external dependencies from orphan check","description":"## Summary\n\nThe `checkForeignKeys` function in `migration_invariants.go` flags external dependencies as orphaned because they dont exist in the local issues table.\n\n## Location\n\n`internal/storage/sqlite/migration_invariants.go` around line 150-162\n\n## Current Code (buggy)\n\n```go\n// Check for orphaned dependencies\nvar orphanCount int\nerr = db.QueryRowContext(ctx, \\`\n SELECT COUNT(*)\n FROM dependencies d\n WHERE NOT EXISTS (SELECT 1 FROM issues WHERE id = d.depends_on_id)\n\\`).Scan(\u0026orphanCount)\n```\n\n## Fix\n\nExclude external references (format: `external:\u003cproject\u003e:\u003ccapability\u003e`):\n\n```go\n// Check for orphaned dependencies (excluding external refs)\nvar orphanCount int\nerr = db.QueryRowContext(ctx, \\`\n SELECT COUNT(*)\n FROM dependencies d\n WHERE NOT EXISTS (SELECT 1 FROM issues WHERE id = d.depends_on_id)\n AND d.depends_on_id NOT LIKE 'external:%'\n\\`).Scan(\u0026orphanCount)\n```\n\n## Reproduction\n\n```bash\n# Add external dependency\nbd dep add bd-xxx external:other-project:some-capability\n\n# Try to sync - fails\nbd sync\n# Error: found 1 orphaned dependencies\n```\n\n## Files to Modify\n\n1. **internal/storage/sqlite/migration_invariants.go** - Add WHERE clause\n\n## Testing\n\n```bash\n# Create issue with external dep\nbd create \"Test external deps\" -t task\nbd dep add bd-xxx external:beads:release-workflow\n\n# Sync should succeed\nbd sync\n\n# Verify dep exists\nbd show bd-xxx --json | jq .dependencies\n```\n\n## Success Criteria\n- External dependencies dont trigger orphan check\n- `bd sync` succeeds with external deps\n- Regular orphan detection still works for internal deps","status":"open","priority":2,"issue_type":"bug","created_at":"2025-12-23T12:37:08.99387-08:00","updated_at":"2025-12-23T12:39:58.536353-08:00"} {"id":"bd-udsi","title":"Async Gates for Agent Coordination","description":"Agents need an async primitive for waiting on external events (CI completion, API responses, human approval). Gates are wisp issues that block until external conditions are met, managed by the Deacon.\n\n## Core Concepts\n\n**Gate** = wisp issue that blocks until external condition is met\n- Type: gate\n- Phase: wisp (never synced, ephemeral)\n- Assignee: deacon/ (Deacon monitors it)\n- Fields: await_type, await_id, timeout, waiters[]\n\n**Await Types:**\n- gh:run:\u003cid\u003e - GitHub Actions run completion\n- gh:pr:\u003cid\u003e - PR merged/closed\n- timer:\u003cduration\u003e - Simple delay\n- human:\u003cprompt\u003e - Human approval required\n- mail:\u003cpattern\u003e - Wait for mail matching pattern\n\n## Open Questions\n- Should gates live in wisp storage or main storage with wisp flag?\n- Do we need a gate catalog (like molecule catalog)?\n- Should waits-for dep type work with gates?","status":"closed","priority":1,"issue_type":"epic","created_at":"2025-12-23T11:44:02.711062-08:00","updated_at":"2025-12-23T12:24:43.537615-08:00","closed_at":"2025-12-23T12:24:43.537615-08:00","close_reason":"Core beads implementation complete (types, migration, CLI). Remaining Deacon integration work moved to gastown (gt-dh65, gt-ng6g, gt-fqcz, gt-gswn)."} {"id":"bd-umbf","title":"Design contributor namespace isolation for beads pollution prevention","description":"## Problem\n\nWhen contributors work on beads-the-project using beads-the-tool, their personal work-tracking issues leak into PRs. The .beads/issues.jsonl is intentionally tracked (it's the project's issue database), but contributors' local issues pollute the diff.\n\nThis is a recursion problem unique to self-hosting projects.\n\n## Possible Solutions to Explore\n\n1. **Contributor namespaces** - Each contributor gets a private prefix (e.g., `bd-steve-xxxx`) that's gitignored or filtered\n2. **Separate database** - Contributors use BEADS_DIR pointing elsewhere for personal tracking\n3. **Issue ownership/visibility flags** - Mark issues as \"local-only\" vs \"project\"\n4. **Prefix-based filtering** - Configure which prefixes are committed vs ignored\n\n## Design Considerations\n\n- Should be zero-friction for contributors (no manual setup)\n- Must not break existing workflows\n- Needs to work with sync/collaboration features\n- Consider: what if a \"personal\" issue graduates to \"project\" issue?\n\n## Expansion Needed\n\nThis is a placeholder. Needs detailed design exploration before implementation.","status":"open","priority":2,"issue_type":"feature","created_at":"2025-12-13T18:00:29.638743-08:00","updated_at":"2025-12-13T18:00:41.345673-08:00"} {"id":"bd-uqfn","title":"Work on beads-wkt: Output control parameters for MCP tool...","description":"Work on beads-wkt: Output control parameters for MCP tools (GH#622). Add brief, fields, max_description_length params to ready/list/show. When done, submit MR (not PR) to integration branch for Refinery.","status":"closed","priority":2,"issue_type":"task","created_at":"2025-12-19T22:57:10.675535-08:00","updated_at":"2025-12-20T00:49:51.929271-08:00","closed_at":"2025-12-19T23:28:25.362931-08:00","close_reason":"Implemented output control parameters for MCP tools (GH#622)"}