codex-review
codex-review - Review Changes Before Pushing
Run codex review on changes and wait for the result before pushing.
Usage
# Review changes since a base commit
codex review --base <base-commit>
# Review uncommitted changes
codex review --uncommitted
Workflow
Step 1 — Start the review in the background
codex review --base <base-commit>
Run with run_in_background: true. Note the task ID.
Step 2 — Wait for completion (do NOT kill on timeout)
Use TaskOutput with a generous timeout (e.g. 300000ms / 5 minutes).
If TaskOutput times out, do NOT kill the task. A timeout means you ran out of patience — not that codex is stuck. The task is likely still making progress.
Step 3 — Check if actually stuck
If you suspect it's stuck (not just slow), compare output file size twice:
wc -c <output_file> # note size
# wait 10 seconds
wc -c <output_file> # if size grew, it's still working
Or watch it live:
tail -f <output_file>
Only kill if file size hasn't grown for 30+ seconds.
Step 4 — Act on the verdict
- Safe: proceed with
git push - Issues found: fix them, then push
- Killed / timed out without verdict: re-run or push with a note that review was inconclusive
Rules
- Always run codex review before
git push, never after - Never kill the review just because
TaskOutputtimed out - If review exits non-zero (e.g. code 144 = killed by signal), treat as inconclusive — re-run if possible
More from snomiao/skills
modernize
Modernize JavaScript/TypeScript toolchain to high-performance native alternatives. Use when upgrading build tools, linters, or compilers.
8heal-pr
Fix and heal a GitHub Pull Request by resolving CI/CD failures, addressing review comments, rebasing conflicts, and iterating until all checks pass. Use when a PR has failing checks, unresolved review comments, merge conflicts, or needs automated review cycles with bots like @copilot.
5claude-api
Build apps with the Claude API or Anthropic SDK. TRIGGER when: code imports `anthropic`/`@anthropic-ai/sdk`/`claude_agent_sdk`, or user asks to use Claude API, Anthropic SDKs, or Agent SDK. DO NOT TRIGGER when: code imports `openai`/other AI SDK, general programming, or ML/data-science tasks.
2skill-creator
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
2