Menu
Run the codereview playbook with a named reviewer agent to verify the current finding against its record. Use when the owner says codereview <agent>.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| name | codereview |
| description | Run the codereview playbook with a named reviewer agent to verify the current finding against its record. Use when the owner says codereview <agent>. |
Run the codereview playbook operator: read
.agents/playbooks/codereview.md and follow it to review the current finding
with the reviewer agent named in the request (for example codereview codex).
The named agent is the reviewer harness; it is dispatched headless and
one-shot per the playbook. If the playbook does not exist in this repo, say so
rather than guessing. The playbook is the authoritative definition; this skill
is only a pointer.
Run the git playbook for plain-English delegated git operations (push, reconcile, add-remote, branch-cleanup). Use when the owner says git <operation> or /git <operation>.
Re-ground in this repo's current state and report before making any change. Use when the owner says catchup or asks where things stand.
Record a settled durable decision in .agents/decisions.md and update affected guidance. Use when the owner says decision.
Compare a doc, decision, or guidance claim against repo evidence and fix or report the conflict; also the deliberate state-hygiene pass. Use when the owner says drift.
Fast save-my-place snapshot of .agents/state.md so the next session resumes without chat context. Use when the owner says handoff or a session is wrapping up.
Run the openreview playbook with a named reviewer agent for an unprimed goal-first review of a whole change. Use when the owner says openreview <agent>.
