Skip to content

amanpal3/Dr-SKILLs / .agent/skills/using-codex-cli

using-codex-cli

Unverified

Orchestrates, runs, and integrates OpenAI's Codex CLI for autonomous coding, non-interactive execution, code reviews, and multi-agent workflows. Use when invoking Codex CLI (`codex exec`, `codex review`), configuring Codex sandboxes and models, sharing skills between Antigravity and Codex, or bridging tools.

Using Codex CLI

Orchestrates, runs, and integrates OpenAI's Codex CLI (codex) alongside modern AI coding agents like Antigravity, Claude Code, and Cursor.

When to Use This Skill

Activate this skill when:

  • Executing autonomous or non-interactive code generation via codex exec
  • Running autonomous code reviews using codex review
  • Configuring Codex sandbox modes, approval policies, and models
  • Setting up shared skill directories across Antigravity (.agent) and Codex (.codex)
  • Delegating sub-tasks or background verification passes to Codex CLI

Skill Architecture & Cross-Tool Compatibility

Codex CLI and Antigravity share a compatible SKILL.md specification format:

Discovery Paths

Codex CLI automatically scans and loads skills from:

  1. User Global Skills: ~/.codex/skills/<skill-name>/SKILL.md
  2. Project Local Skills: .codex/skills/, .agents/skills/, or .agent/skills/
  3. Canonical Instructions: CODEX.md and AGENTS.md in the project root

Instruction Hierarchy

In any project repository:

  • CODEX.md / AGENTS.md: Instructs Codex CLI on project architecture, skill loading rules, and constraints.
  • .agent/skills/<skill-name>/SKILL.md: The actionable domain instructions loaded on-demand by Codex CLI.

Codex CLI Execution Recipes

1. Non-Interactive Autonomous Execution (codex exec)

Run a prompt headlessly with controlled workspace write access:

# Execute a task in the current workspace
codex exec -s workspace-write --approve-for-me "Implement user authentication in auth.ts following the owasp-security skill"

Options:

  • -s, --sandbox <MODE>: read-only, workspace-write, danger-full-access
  • --approve-for-me: Automatically reviews and approves safe tool operations
  • -m, --model <MODEL>: Specify model (e.g., gpt-5.6-terra, o3, or local OSS)
  • --search: Enables live web search without per-call confirmation prompts

2. Autonomous Code Review (codex review)

Run an automated review of uncommitted changes or branches:

codex review

3. Applying Diffs (codex apply)

If a task was planned or produced by Codex in a separate session, apply its git patch:

codex apply

4. Background Delegation from Antigravity / Orchestrators

To delegate a specialized sub-task to Codex CLI from Antigravity:

# Run Codex non-interactively in background
codex exec -C "C:\path\to\repo" -s workspace-write --approve-for-me "Run unit tests and fix failing assertions"

Setting Up Codex CLI in Any Project

When initializing or working in a repository with Codex CLI:

1. Add CODEX.md to Project Root

Create a CODEX.md linking Codex to the .agent/skills folder:

# CODEX.md � Instructions for Codex CLI

## Project Overview
This project uses a modular skill architecture located in `.agent/skills/`.

## Skill Loading Rules
- Always check `.agent/skills/<skill-name>/SKILL.md` before starting work in that domain.
- When working on APIs or database schemas, load `building-backend` and `integrating-sentry`.
- When designing UI or screens, load `ui-ux-pro-max` and `designing-with-stitch`.
- When reviewing or refactoring code, load `ponytail` and `code-review`.
- Always handle errors gracefully with Sentry instrumentation.

2. Sync Global Skills

Ensure all 43 skills are synced to ~/.codex/skills/:

.\install-skills.ps1 store

Troubleshooting & Verification

  • Check Codex installation: codex --version
  • Run diagnostic health check: codex doctor
  • List configured MCP servers: codex mcp list
  • List installed plugins: codex plugin list
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 464kUpdated Claude CodeCodexVerifiedInstall
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 464kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 318kUpdated Claude CodeCodexUnverifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 318kUpdated Claude CodeCodexUnverifiedInstall