skills.rest
fghaffar
fghaffarCOMMUNITY
Explore all skills by this author →
GitHub

codex

Review git diffs and plan files with Codex and return gate verdicts.

Updated Mar 25, 2026
Search Tags:#code review#security#cli#codex#diff#adversarial testing#cross-model
One-click install
npx skills add https://github.com/fghaffar/faystack --skill codex-fghaffar
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: codex
Source: https://github.com/fghaffar/faystack/tree/main/codex
Command: npx skills add https://github.com/fghaffar/faystack --skill codex-fghaffar

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It provides an independent, focused AI second opinion on code changes, plans, and prompts so teams catch issues, design flaws, and security gaps before shipping.

Core Features & Use Cases

  • Review Mode: Runs Codex against a git diff to produce a verbatim code review and a pass/fail gate based on critical findings.
  • Challenge Mode: Adversarial testing that tries to break assumptions in code or designs to surface edge cases and failure modes.
  • Consult Mode: Interactive Q&A with session continuity for follow-ups and targeted guidance.
  • Use Case: Run on a feature branch to validate security, correctness, and edge cases before merging, or ask for a focused consult about a tricky algorithm.

Quick Start

Run /codex review to get an independent Codex-driven code review of the current branch diff.

Frequently Asked Questions about codex

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run an AI code review on a git diff before merging?▼

To run an AI code review on a git diff, use the review mode to process your branch changes. It returns verbatim Codex output and a pass/fail gate verdict based on critical security and correctness findings.

Can I use adversarial testing to break assumptions in my code designs?▼

Yes, you can use adversarial testing through the challenge mode to break assumptions in your code designs. It actively surfaces edge cases and failure modes in repository diffs and local plan files.

What is the best way to get an independent second opinion on a local plan file?▼

The best way to get an independent second opinion on a local plan file is using consult mode. It provides interactive Q&A with session continuity for targeted guidance and follow-up questions.

Does the codex skill access files outside my git repository?▼

No, the codex skill enforces a strict filesystem boundary that excludes ~/.claude and .claude/skills directories. It processes repository diffs and plan files while respecting this security constraint.

How do I increase the strictness of an AI code review for critical security gaps?▼

You can increase the strictness of an AI code review by applying the --xhigh mode flag. This elevates the analysis depth when evaluating branch diffs and direct consult prompts.

Related Skills

Explore More Software Engineering Skills→
RheasilviaRheasilvia

github-code-review

Analyze git diffs and submit structured GitHub pull request reviews.

Community
Advanced
BillChiricoBillChirico
5

github-pr-resolver

Resolve GitHub pull request review threads and failing CI checks with commits.

Community
Advanced
singhianand23045singhianand23045

codex

Automate Codex CLI code review, adversarial testing, and consult workflows.

Community
Advanced
uphiagouphiago
1.1k

hunt-llm-ai

Identify exploitable LLM/AI vulnerabilities crossing trust boundaries with validation gates.

Community
Advanced