Skip to content

OpenAI Codex Skills — Setup Guide

Source: openai/skills (38K+ combined installs) GitHub: openai/skills (24K+ ⭐) Category: Agent Infrastructure / Coding Agents Quality Tier: 🟢 Production

OpenAI's official skills catalog for Codex CLI — their agentic coding tool. These skills extend Codex with PDF generation, GitHub CI/CD automation, security best practices, Linear issue tracking integration, Playwright browser testing, and Figma-to-code design implementation. For Hermes agents that use Codex as an implementation backend, these skills add production-ready capabilities.


Installation

# Document generation
npx skills add openai/skills --skill pdf

# CI/CD + GitHub automation
npx skills add openai/skills --skill gh-fix-ci

# Security
npx skills add openai/skills --skill security-best-practices

# Issue tracking + testing
npx skills add openai/skills --skill linear
npx skills add openai/skills --skill playwright

# Design implementation
npx skills add openai/skills --skill figma-implement-design

Included Skills

Skill Installs Purpose
pdf 10.4K Generate, parse, and manipulate PDF documents programmatically
gh-fix-ci 7.8K Diagnose and fix GitHub Actions CI/CD pipeline failures
security-best-practices 5.7K Security audit workflows — dependency scanning, secret detection, OWASP patterns
linear 5.2K Linear issue tracking integration — create, update, and query issues
playwright 4.7K Browser automation testing with Playwright — E2E tests, visual regression
figma-implement-design 4.4K Convert Figma designs to production code with component matching

Prerequisites

Requirement Details
Codex CLI npm install -g @openai/codex (or npx @openai/codex)
OpenAI API key Required for Codex agent usage
Node.js 18+ Required for Codex runtime

Key Capabilities

Document Generation (pdf)

Generate professional PDF reports, invoices, and documentation directly from code. Supports PDF parsing for data extraction and template-based generation.

CI/CD Automation (gh-fix-ci)

Intelligently diagnose GitHub Actions failures by reading workflow logs, identifying root causes, and proposing fixes. Handles common failure patterns: dependency resolution, environment variables, artifact uploads.

Security Auditing (security-best-practices)

Automated security review workflows covering OWASP Top 10, dependency vulnerability scanning (npm audit, pip audit), secret detection, and secure coding patterns.

Browser Testing (playwright)

Write and execute Playwright tests for web applications. Includes visual regression testing, accessibility audits, and multi-browser coverage.

Design-to-Code (figma-implement-design)

Parse Figma design files and generate matching React/Vue components with accurate styling, spacing, and responsive breakpoints.


Quick Start

# 1. Install Codex CLI
npm install -g @openai/codex

# 2. Add the CI fix skill
npx skills add openai/skills --skill gh-fix-ci

# 3. Fix a broken CI pipeline
codex "Using the gh-fix-ci skill, diagnose why the main branch CI is failing"

# 4. Generate a PDF report
npx skills add openai/skills --skill pdf
codex "Using the pdf skill, generate a sprint review report from the last week of commits"

Verification

# Check installed OpenAI skills
npx skills list | grep openai/skills

# Verify Codex can use them
codex "list available skills" 2>&1 | grep -i openai

Notes

  • These are the Codex-equivalent of Anthropic's Claude Code skills — both extend their respective agentic coding tools
  • The gh-fix-ci skill is particularly valuable for Hermes agents managing GitHub Actions across multiple repos
  • figma-implement-design requires Figma API access token for private design files
  • For Hermes agents using Codex as an implementation backend, add the full skill set to unlock PDF, CI, security, and testing capabilities
  • Distinct from anthropics/claude-code skills — both catalogs serve the same purpose (agent extensibility) but for different agent runtimes