The previous bd-in7q fix had backwards logic - by EXCLUDING tombstones
from currentIDs, they appeared missing when compared to historicalIDs,
causing HydrateDeletionsManifest to erroneously add them to deletions.jsonl.
This corruption manifested when:
1. Issues were migrated to tombstones via migrate-tombstones
2. Doctor hydration ran (directly or via sync)
3. Tombstones were seen as deleted and re-added to deletions.jsonl
4. Next import skipped thousands of issues with in deletions manifest
Fix: Include ALL issues (including tombstones) in currentIDs. Tombstones
represent migrated deletions that ARE accounted for - they should not
trigger new deletion records.
Generated with Claude Code
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add --stop-all flag to bd daemon command that:
- Discovers all running bd daemon processes using the registry
- Gracefully stops them all (with force kill fallback)
- Reports how many were stopped
Useful for:
- Cleaning up multiple daemons causing race conditions
- Getting a clean slate before running bd sync
- Debugging daemon-related issues
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The orphan detection was incorrectly flagging issues with dots in their
prefix (e.g., "my.project-abc123") as orphans because it was looking for
any dot in the ID, treating everything before the first dot as the
parent ID.
The fix:
- Add IsHierarchicalID() helper that correctly detects hierarchical IDs
by checking if the ID ends with .{digits} (e.g., "bd-abc.1")
- Update SQL query in orphan detection migration to use GLOB patterns
that only match IDs ending with numeric suffixes
- Update all Go code that checks for hierarchical IDs to use the new
helper function
Test cases added:
- Unit tests for IsHierarchicalID covering normal, dotted prefix, and
edge cases
- Integration test verifying dotted prefixes do not trigger false
positives
Fixes: #508
Allows changing issue type (task/epic/bug/feature/chore) via bd update --type.
Storage layer already supported it, this adds CLI and RPC support.
Fixes GH#522.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Fixes GH#523: older versions of bd could close issues without setting
the closed_at timestamp. When importing such issues, validation would
fail with "closed issues must have closed_at timestamp".
This fix adds defensive handling in all issue creation/validation paths:
- If status is "closed" and closed_at is nil, set closed_at to
max(created_at, updated_at) + 1 second
- Similarly for tombstones missing deleted_at
Applied to:
- batch_ops.go: validateBatchIssuesWithCustomStatuses (main import path)
- transaction.go: CreateIssue and CreateIssues
- queries.go: CreateIssue
- multirepo.go: upsertIssueInTx
Also adds comprehensive tests for the defensive fix.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add support for claude.local.md and .claude/claude.local.md as local-only
alternatives to CLAUDE.md. These files are intended to be gitignored for
personal customizations that should not be committed to the repo.
Changes:
- Update CheckAgentDocumentation to detect claude.local.md files
- Update CheckLegacyBeadsSlashCommands to check claude.local.md files
- Update CheckDocumentationBdPrimeReference to check claude.local.md files
- Add fix message mentioning claude.local.md as an option
- Add test cases for claude.local.md support
Closes#488🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When sync.branch is configured to the same branch as the current branch,
git worktree creation fails because the same branch cannot be checked out
in multiple locations.
This fix detects when sync.branch equals the current branch and falls back
to direct commits on the current branch instead of using the worktree-based
approach.
Changes:
- Add IsSyncBranchSameAsCurrent() helper in syncbranch package
- Add GetCurrentBranch() helper function
- Update sync.go to detect this case and skip worktree operations
- Add unit tests for the new functionality
Closes#519🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When worktrees are nested under the main repo (e.g.,
/project/.worktrees/feature/), bd now correctly finds .beads/ in the
parent repo.
The fix simplifies GetMainRepoRoot() to use `git rev-parse --git-common-dir`
which always returns the main repo's .git directory, regardless of whether
we're in a regular repo, a worktree, or a nested worktree.
- Simplified GetMainRepoRoot() implementation
- Added tests for nested worktree scenarios
- Updated CHANGELOG.md
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Closes GH#516: bd init now automatically:
- Creates AGENTS.md and @AGENTS.md with landing-the-plane instructions if they don't exist
- Appends the instructions to existing files if they don't have them
- Skips if the section already exists (idempotent)
- Skips in stealth mode (user wants invisible setup)
The landing-the-plane instructions ensure AI agents properly complete
their work sessions by pushing all changes to remote before ending.
Fixes GH#483 - The pre-commit hook was blocking commits when
`bd sync --flush-only` failed, even if the user had removed beads
from their branch. This made it impossible to commit on branches
that don't have beads configured.
Changes:
- Change "Error:" to "Warning:" in the message
- Remove `exit 1` so commits proceed even if flush fails
- Add comments explaining why we don't block commits
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add comprehensive documentation about beads-created worktrees:
- New section in WORKTREES.md explaining why and where beads creates worktrees
- Common confusion: 'beads took over main!' with solution
- Troubleshooting entries for 'branch already checked out' errors
- FAQ entry: 'Why did beads create worktrees in my .git directory?'
This addresses user confusion when beads' sync-branch feature creates
internal worktrees that can lock branches users want to checkout.
Closes#510🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add comprehensive documentation for the `bd setup` command:
- New docs/SETUP.md with full documentation for all three integrations:
- Claude Code (hooks for SessionStart/PreCompact)
- Cursor IDE (rules file in .cursor/rules/)
- Aider (config file with human-in-the-loop workflow)
- Updated docs/CLI_REFERENCE.md:
- Added Setup & Integration section
- Added SETUP.md to See Also links
- Added to Quick Navigation
Closes#518🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Update from v0.24.2 to v0.29.0 to get the generic template content
(removes beads-specific Go/SQLite/Cobra content).
Relates to: GH#497
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
GH#517: Claude uses 'medium' instead of P2/2 for priority, causing
infinite error loops. The bd prime output didn't show the --priority
flag or explain the valid format.
Fix: Add --priority=2 to the create example and add a clear note
explaining that priority must be 0-4 or P0-P4, NOT 'high/medium/low'.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
These entries (beads-, test-, webui-, worker2-) were hydrated from git
history but don't belong in this repo's bd- namespace. They were causing
import failures during beads sync due to prefix mismatch validation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
