- use-skills.sh: symlink to $CODEX_HOME/skills when CODEX_HOME is set
- docs: update PER-REPO-SKILLS.md and RFC-SKILLS-MANIFEST.md with Codex flow
- hq: add model configuration section (sonnet-4.5, Claude Max)
- hq: update launch commands with explicit --model flag
Closes skills-legi
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add codexSkills option for deploying skills to ~/.codex/skills/
- Follows same pattern as claudeCodeSkills and openCodeSkills
- Dotfiles can now configure: services.ai-skills.codexSkills = [ "worklog" "hq" ];
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add try/except blocks in worker commands to catch WorkerNotFound,
InvalidTransition, and StaleState exceptions
- Return ExitInvalidTransition (3) for state transition errors
- Return ExitNotFound (7) for missing workers
- Fix double ROLLBACK bug in state.nim by removing inline ROLLBACKs
and letting the except block handle transaction cleanup
Closes: skills-lxb9
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Test infrastructure:
- Add comprehensive test-worker.sh with 59 tests
- Cover spawn, start, done, approve, reject, cancel, fail, heartbeat
- Test status filtering, context files, review-gate integration
- Test invalid state transitions and error cases
Bug fixes:
- Fix agent commands failing when run from worktrees (skills-y3f2)
- Add getMainRepoBusDbPath() to find DB in main repo
- Update start, done, fail, heartbeat to use correct path
- Fix review.nim crash when review-gate not installed (OSError)
- Gracefully return exitCode=-1 instead of crashing
Closes skills-y3f2
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Documents rationale for using Nim with ORC for the worker coordination
CLI: fast startup, single binary, Python-like syntax, excellent SQLite
support via tiny_sqlite, CLI generation via cligen.
Closes skills-q40
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- CLI tool with check/enable/approve/reject/status commands
- Stop hook integration for Claude Code blocking
- Adversarial reviewer prompt template
- Dual-publish structure (SKILL.md + .claude-plugin)
- Session ID auto-detection from env vars or git state
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Comprehensive comparison of Claude Code, Gemini CLI, OpenCode, and Codex:
- Hooks/lifecycle events (Claude/Gemini best, OpenCode most comprehensive)
- Subagent spawning (MCP is universal bridge)
- File access (Gemini has path restrictions - skills-bo8)
- Sandboxing (Codex has OS-level, others approval-based)
- State persistence (need external store for cross-agent)
Key finding: Orchestrator pattern works across all agents.
Stop hooks only in Claude/Gemini - others need protocol-based gates.
Closes: skills-fqu
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
