skill-code-review

github:ctxr-dev/skill-code-review · first committed 2026-04-11 · source · badge

Use when completing tasks, implementing major features, or before merging to verify work meets requirements. Runs an FSM-driven, deterministic, manifest-producing code-review pipeline via scripts/run-review.mjs. The runner stages every intermediate (brief, agent dispatch prompt, worker output) under <run_dir>/workers/. The orchestrator drives a shell loop using --print-run-dir / --print-current-st

Skill Health: not yet scored

Listed, not yet scored. This skill has no public model test yet. It is in the scoring queue; submitted skills go first, then the most-installed.

Version history

21 commits to this file on GitHub (newest first).

2026-06-08 · commit cc3377d · refactor: code_review rename + deterministic-wiki layout/validators + tmp re-shard
2026-06-07 · commit 75034ba · feat(runner): fault-tolerant ThreadPoolExecutor runner for the FSM
2026-06-06 · commit 0c6fde7 · fix(specialists): hard-raise DispatchLoopExitedEarlyError when dispatch loop drained early (#114)
2026-06-03 · commit 5058b96 · feat(skill): verifier_stuck retry cap + aggregate-size guard + observability docs (#112)
2026-06-03 · commit f5b25a7 · feat(skill): pin allowed_tools per worker state + tool-surface table in SKILL.md (#109)
2026-06-02 · commit 5281497 · docs(skill): require user consent before any package-missing install (#107)
2026-06-01 · commit 6e76ab0 · feat: declare prompt_template_language='markdown' on all 5 workers (#106)
2026-05-22 · commit f29ce1c · feat: cross-harness dispatch (subagent.dispatch.v1 + Agent→sub-agent prose) (#104)
2026-05-02 · commit a49b5a7 · fix: skill works when installed outside its own repo (closes #100) (#101)
2026-05-02 · commit 13b45f3 · feat(skill): fatter shards + --print-batch-envelope CLI for fewer Agents and shorter loops (#95)
2026-05-01 · commit 6255d60 · feat(skill): per-leaf specialist outputs + diff sharding + 10-agent batched dispatch + shim docs (#93)
2026-05-01 · commit 08fa397 · feat(skill): per-leaf file_globs + pre-computed filtered diffs (closes #83) (#84)
2026-04-30 · commit 2b37464 · feat(skill): pre-stage agent prompts under run-dir; eliminate /tmp from orchestrator (closes #79) (#80)
2026-04-29 · commit 5e7a22a · feat(skill): run-dir as the source of truth (closes #76) (#77)
2026-04-29 · commit f0e54d7 · refactor(skill): orchestrator dispatches specialists directly (PR D of #70) (#74)
2026-04-29 · commit be3524a · fix(skill): tighten verbatim-surface, fault-stop, and prompt body in brief (#70) (#71)
2026-04-29 · commit c56cd50 · feat(skill): structurally enforce run-review.mjs invocation (closes #64) (#65)
2026-04-26 · commit 87f0a2a · docs: linear step numbering, strip historical framing, FSM as P0 Sprint A
2026-04-25 · commit 9f1823e · refactor: orchestrate against reviewers.wiki tree (Phase 1.Z + 2 + 3)
2026-04-11 · commit 7122909 · feat: add performance, README quality, release readiness, security, and test quality reviewers
2026-04-11 · commit 1912a9d · Add comprehensive review guidelines for release readiness, security, and test quality

Archived snapshots