reqCtx() now applies the server's requestTimeout (default 30s) to the
context returned for request handlers. This prevents bd list and other
commands from hanging indefinitely when database operations stall.
The fix ensures:
- All RPC handlers get a context with a deadline
- Database operations (using QueryContext) honor context cancellation
- reconnectMu read locks are released when context times out
Fixes: bd-p76kv
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Same fix as PR #991 for FindBeadsDir() - the loop condition
dir != "/" && dir != "." doesn't handle Windows drive roots.
On Windows, filepath.Dir("C:\\") returns "C:\\", not "/" or ".".
Changed both functions to check parent == dir to detect filesystem root,
which works correctly on both Unix and Windows.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Insert git config user.name in fallback chain before final $USER default
- Consolidate duplicate actor resolution logic into single function
- Add BEADS_ACTOR as env var alias for MCP compatibility
- Add tests for actor resolution priority
- Update CONFIG.md with Actor Identity Resolution section
The new fallback order is:
--actor flag > BD_ACTOR > BEADS_ACTOR > git user.name > $USER > "unknown"
Co-authored-by: Ohffs <ohffsnotnow@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
On Unix systems, socket paths are limited to 104 chars (macOS) or 108 chars
(Linux). Deep workspace paths like /Volumes/External Drive/Dropbox/...
would exceed this limit and cause daemon startup failures.
This fix:
- Adds ShortSocketPath() which computes /tmp/beads-{hash}/bd.sock for
paths that would exceed the limit
- Keeps backward compatibility: short paths still use .beads/bd.sock
- Updates daemon discovery to check both locations
- Uses SHA256 hash of canonical workspace path for unique directories
Closes GH#1001
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
warnIfCyclesExist was called with nil storage when using daemon mode.
Add nil check to skip cycle warning in daemon mode (daemon handles it).
Fixes#998
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds ability to burn multiple molecules in a single command:
bd mol burn bd-a1 bd-b2 bd-c3
Changes:
- Accept multiple molecule IDs as arguments
- Add BatchBurnResult struct for aggregated results
- Implement burnMultipleMolecules for batch efficiency
- Wisps deleted in single batch call
- Maintains backward compatibility for single ID
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Use SearchIssues with filter.IDs instead of GetIssue in ResolvePartialID.
This fixes inconsistencies where bd list could find issues that bd show
could not, due to subtle query differences between the two code paths.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When removing all hooks from an event, the key was being set to null
instead of being deleted. Claude Code expects arrays, not null values,
causing startup failures with 'Expected array, but received null'.
Changes:
- removeHookCommand now deletes the event key when no hooks remain
- installClaude cleans up any existing null values from buggy removal
- Added tests for null value prevention and cleanup
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
On Windows, the daemon stop command was failing with "exit status 1" because
`taskkill` without `/F` flag does not work for console processes that do not
have windows to receive close messages.
Changes:
- Use `os.Process.Kill()` which calls Windows `TerminateProcess` API
- This is the reliable way to terminate processes on Windows
- The graceful RPC shutdown is already attempted before falling back to kill
- Updated error messages to be platform-agnostic (removed SIGTERM/SIGKILL)
The fix uses Go cross-platform process APIs instead of shelling out to
external commands, which is more reliable and portable.
Closes#992
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Auto-detect WSL2 environment with Windows filesystem paths (/mnt/c/, etc.)
and fall back to DELETE journal mode instead of WAL. SQLite WAL mode
does not work reliably across the WSL2/Windows 9P filesystem boundary
due to shared-memory limitations.
Detection checks:
1. Path matches /mnt/[a-z]/ pattern (Windows drive mount)
2. /proc/version contains microsoft or wsl
Fixes#920
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
The sync command was closing the daemon connection without initializing
the direct store, leaving store=nil. This caused errors in post-checkout
hook when running bd sync --import-only.
Fixed by using fallbackToDirectMode() which properly closes daemon and
initializes the store.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Move all plugin content to claude-plugin/ subdirectory for cleaner
separation from core beads functionality:
- claude-plugin/.claude-plugin/plugin.json - plugin manifest
- claude-plugin/commands/ - all slash commands
- claude-plugin/agents/ - task-agent (now in correct location)
- claude-plugin/skills/beads/ - beads skill with resources
Root .claude-plugin/ now only contains marketplace.json pointing
to ./claude-plugin.
Note: After updating, users may need to clear plugin cache:
rm -rf ~/.claude/plugins/cache/beads-marketplace/
Fixes#985
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Root cause: CreateIssue used INSERT OR IGNORE which could silently skip
the insert (e.g., on duplicate ID from hash collision), then fail with
FOREIGN KEY constraint error when trying to record the creation event.
Fix: Add insertIssueStrict() that uses plain INSERT (fails on duplicates)
and use it for CreateIssue in both queries.go and transaction.go. The
existing insertIssue() with INSERT OR IGNORE is preserved for import
scenarios where duplicates are expected.
Added test TestCreateIssueDuplicateID to verify duplicate IDs are properly
rejected instead of silently ignored.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Added IsBlocked method to Storage interface that checks if an issue is
in the blocked_issues_cache and returns the blocking issue IDs.
The close command now checks for blockers before allowing an issue to
be closed:
- If an issue has open blockers, closing is blocked with an error message
- The --force flag overrides this check
- Works in both daemon mode (RPC) and direct storage mode
- Also handles cross-rig routed IDs
This addresses the bug where agents could close a bead even when it
depends on an open bug/issue.
Closes#962
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add support for lefthook v1.10.0+ jobs array syntax in addition to
the older commands map syntax. Both syntaxes are now checked for
bd hooks run integration.
Fixes#981
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Parse $EDITOR value to handle editors that need flags like
"zeditor --wait" or "code --wait". Previously the entire string
was treated as the executable name.
Fixes#987
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
v0.46.0 removed these types breaking gt install/doctor/sling/convoy commands.
This restores them as built-in types so `bd create --type=agent` works again.
Fixes GH#941
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds a diagnostic check that discovers and reports custom types
used by child repos in multi-repo setups. This is the 'discover'
part of the 'trust + discover' pattern for federation.
The check:
- Lists custom types found in each child repo's config/DB
- Warns about hydrated issues using types not found anywhere
- Is informational only (doesn't fail overall doctor check)
Part of epic: bd-9ji4z
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Implements federation trust model for multi-repo type validation:
- Built-in types are validated (catch typos)
- Non-built-in types trusted from source repos
Changes:
- Add IssueType.IsBuiltIn() method
- Add Issue.ValidateForImport() for trust-based validation
- Update upsertIssueInTx to use ValidateForImport
Closes: bd-dqwuf, bd-alpw2 | Epic: bd-9ji4z
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When database is stale and a read-only command runs in --no-daemon mode,
auto-import from JSONL instead of returning an error. This fixes `gt sling`
failing with 'not a valid bead' after git pull.
Root cause: `ensureDatabaseFresh` would return an error when DB was stale,
but read-only commands like `bd show` should be able to auto-import and
proceed rather than blocking the user.
Changes:
- Modify ensureDatabaseFresh to call autoImportIfNewer() when stale
- Only error if --no-auto-import flag is explicitly set
- Add comprehensive tests for all staleness scenarios
Fixes: bd-9dao
Co-authored-by: Claude <noreply@anthropic.com>
After devcontainer restart (cold-start), `bd --no-daemon show` failed to
find beads because:
1. Read-only commands skipped auto-import
2. Newly created DB had no issue_prefix set, causing import to fail
This fix enables seamless cold-start recovery by:
- Allowing read-only commands (show, list, etc.) to auto-bootstrap when
JSONL exists but DB doesn't
- Setting needsBootstrap flag when falling back from read-only to
read-write mode for missing DB
- Auto-detecting and setting issue_prefix from JSONL during auto-import
when DB is uninitialized
Fixes: gt-b09
Co-authored-by: Claude <noreply@anthropic.com>
* fix(doctor): add .sync.lock and sync_base.jsonl to gitignore
Problem:
- .sync.lock and sync_base.jsonl were missing from GitignoreTemplate
- Files introduced in PR #918 (pull-first sync) appeared as untracked
Solution:
- Add patterns to GitignoreTemplate with explanatory comment
- Add patterns to requiredPatterns for bd doctor validation
Impact:
- Existing users get warning via bd doctor, fix via bd doctor --fix
- New repos get correct gitignore on bd init
Fixes: GH#974
* docs(sync): fix sync_base.jsonl tracking status and add tests
Problem:
- SYNC.md incorrectly documented sync_base.jsonl as "git-tracked"
- No tests validated sync state file gitignore patterns
Solution:
- Update Files Reference table: sync_base.jsonl is "not tracked, per-machine"
- Add TestGitignoreTemplate_ContainsSyncStateFiles
- Add TestRequiredPatterns_ContainsSyncStatePatterns
Relates to: GH#974
Two issues caused `bd migrate sync` to fail when run from a git worktree:
1. Used GetGitDir() instead of GetGitCommonDir() for worktree path
- GetGitDir() returns the worktree-specific path (.bare/worktrees/main)
- GetGitCommonDir() returns the shared git dir (.bare) where new
worktrees can actually be created
2. Used strings.Index instead of LastIndex in GetRepoRoot()
- When user paths contain "worktrees" (e.g., ~/Development/worktrees/),
Index finds the first occurrence and incorrectly strips the path
- LastIndex finds git's internal /worktrees/ directory
Added GetGitCommonDir() to internal/git/gitdir.go for reuse.
Fixes GH#639 (remaining unfixed callsite in migrate_sync.go)
When running `bd sync --import-only` while the daemon is connected, the command
fails with "no database store available for inline import".
Root cause:
1. PersistentPreRun connects to daemon and returns early without initializing
the store global
2. sync command closes the daemon connection (for consistency)
3. sync --import-only calls importFromJSONLInline which requires store != nil
4. Without ensureStoreActive(), the store is never initialized after daemon disconnect
Fix: Call ensureStoreActive() after closing the daemon connection in sync.go.
This ensures the local SQLite store is initialized for all sync operations
that need direct database access.
- Add ensureStoreActive() call after daemon disconnect in sync.go
- Add test documenting the bug and verifying the fix
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add AllowStale field to ListArgs struct to support resilient hook detection.
When set, callers signal they accept potentially stale data rather than
failing on staleness check errors.
This enables gastown checkSlungWork() to fall back gracefully when the
beads database is out of sync with JSONL (common after concurrent agent syncs).
- Add AllowStale bool to ListArgs in internal/rpc/protocol.go
- Pass --allow-stale flag through to RPC in cmd/bd/list.go
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
Add new command to find molecules where a gate has closed and the workflow
is ready to resume:
- `bd ready --gated` - Flag on existing ready command
- `bd mol ready` - Subcommand for discoverability
The command finds molecules where:
1. A gate bead (type=gate) has been closed
2. The step blocked by that gate is now ready
3. The molecule is not currently hooked by any agent
This enables the Deacon patrol to discover and dispatch gate-ready
molecules without explicit waiter tracking, supporting async molecule
resume workflows.
Includes 5 tests verifying:
- No false positives when no gates exist
- Detection of molecules with closed gates
- Filtering out molecules with open gates
- Filtering out already-hooked molecules
- Handling multiple gate-ready molecules
Part of epic bd-ka761 (Gate-based async molecule resume).
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Executed-By: beads/crew/dave
Rig: beads
Role: crew
bd list --tree:
- Use actual parent-child dependencies instead of dotted ID hierarchy
- Treat epic dependencies as parent-child relationships
- Sort children by priority (P0 first)
- Fix tree display in daemon mode with read-only store access
bd graph:
- Add --all flag to show dependency graph of all open issues
- Add --compact flag for tree-style rendering (reduces 44+ lines to 13)
- Fix "needs:N" cognitive noise by using semantic colors
- Add blocks:N indicator with semantic red coloring
bd show:
- Tufte-aligned header with status icon, priority, and type badges
- Add glamour markdown rendering with auto light/dark mode detection
- Cap markdown line width at 100 chars for readability
- Mute entire row for closed dependencies (work done, no attention needed)
Design system:
- Add shared status icons (○ ◐ ● ✓ ❄) with semantic colors
- Implement priority colors: P0 red, P1 orange, P2 muted gold, P3-P4 neutral
- Add TrueColor profile for distinct hex color rendering
- Type badges for epic (purple) and bug (red)
Design principles:
- Semantic colors only for actionable items
- Closed items fade (muted gray)
- Icons > text labels for better scanability
Co-Authored-By: SageOx <ox@sageox.ai>
- resolve_conflicts.go: Mark unused `num` parameter with `_`
- .golangci.yml: Add resolve_conflicts.go to G306 exclusion (JSONL files use 0644)
- .golangci.yml: Add doctor/git.go to G304 exclusion (safe path construction)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Problem:
- When dbPath is set to relative fallback (".beads/beads.db"),
findJSONLPath() returns a relative path
- daemon_sync_branch.go calls filepath.Rel(absolutePath, relativePath)
which fails with: "Rel: can't make .beads/issues.jsonl relative to ..."
Solution:
- Canonicalize dbPath at assignment in main.go:471 (source fix)
- Add defensive guard in findJSONLPath() (defense-in-depth)
- Use utils.CanonicalizePath() for OS-agnostic handling
(symlinks, case normalization on macOS)
Testing:
- Add TestFindJSONLPath_RelativeDbPath (tracer bullet for bug)
- Add edge case tests for BEADS_JSONL and empty dbPath
- All sync mode tests pass including daemon E2E
Reviewed by beads/crew/wolf. Fixes daemon mode silently ignoring --due and --defer flags. Adds comprehensive tests including TestDualPathParity for regression prevention.
