2btwist/dotagents/codex-review

Claude Code and agent skill by 2btwist/dotagents

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Get an independent code review from Codex CLI (the OpenAI coding agent) of uncommitted changes, a branch diff, or a commit, as a second perspective alongside Claude's own review. Use when the user asks for a Codex/GPT review or a second opinion on a diff, or before shipping broad or risky changes where an independent reviewer counters author bias. Claude verifies Codex's findings against the code before relaying them.

Install

npx vskill@latest install 2btwist/dotagents/codex-review

SKILL.md

The same SKILL.md was published earlier as 2btwist/dotagents/codex-review.

Codex review

Adapted from Theo's (t3.gg) setup. An independent reviewer that did not author the change; useful precisely because of the author-bias rule (never be the sole verifier of your own output). Treat its output as evidence, not authority.

Prefer Claude's normal review for small local checks. Do not delegate review just to avoid reading the code.

Command shapes

Codex CLI (>= 0.144.x) treats a custom PROMPT (positional or - stdin) as mutually exclusive with every review target flag (--commit, --base, --uncommitted), despite what its usage line suggests. Run the target flags bare; carry context another way (see below).

bash
ARTIFACT_DIR="$(mktemp -d "${TMPDIR:-/tmp}/codex-review.XXXXXX")"
# uncommitted (staged + unstaged + untracked)
codex -C "$PWD" review --uncommitted > "$ARTIFACT_DIR/report.md"
# branch vs base
codex -C "$PWD" review --base main > "$ARTIFACT_DIR/report.md"
# single commit
codex -C "$PWD" review --commit <sha> > "$ARTIFACT_DIR/report.md"

Long diffs take minutes: pass an explicit Bash timeout (600000) or run in the background.

Getting context into the review

With target flags, Codex sees only its default instructions plus the diff and repo:

  • A detailed commit message is the main context channel for --commit reviews; write invariants and risky areas into it before requesting review.
  • For custom instructions, use codex exec with a self-contained prompt (name the files/sha to inspect, the invariants to check, and "do not edit files"); it costs more quota than review.

Reporting back

Verify each important finding against the code before relaying it; separate confirmed issues from unverified Codex suggestions. If Codex finds nothing, say that and name the review target it inspected. If codex is missing or fails, review directly instead.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/2btwist/dotagents/codex-review/badge)](https://verified-skill.com/skills/2btwist/dotagents/codex-review)

Something wrong with this skill or its page? Tell us.