A skill is a set of instructions packaged as a SKILL.md file that an AI agent reads to learn a new workflow. Codex CLI, OpenAI's command-line coding agent, supports the SKILL.md format alongside its own openai.yaml metadata file, making it versatile and easy to extend with community or custom workflows.
What Are Skills in Codex CLI?
Skills are essentially modular, reusable workflows or capabilities that augment the Codex CLI's functionality. Packaged as SKILL.md files, these skills instruct the AI agent on how to perform specific tasks, ranging from code generation, formatting, testing, to deployment scripts. This modularity allows developers to customize their AI assistant to fit their unique development environment and workflow requirements.Before You Start: Prerequisites
Before installing skills, ensure that you have Codex CLI installed and functioning correctly on your system. You can verify this by running:```bash codex --version ```
If this command returns a version number, you’re good to go. If not, install Codex CLI from the official source or via your package manager. Additionally, skills are stored in a specific directory, typically at `~/.codex/skills/`. If this directory doesn’t exist, create it manually with:
```bash mkdir -p ~/.codex/skills/ ```
This directory acts as the central repository for all your SKILL.md files, making management easier.
Step-by-Step Guide: Installing Skills in Codex CLI
Follow these steps to install a new skill from Agensi, GitHub, or your own project repositories. 1. Download the SkillVisit Agensi Skills or a GitHub repository containing SKILL.md files. Download the skill files, usually packaged as zip archives or individual markdown files. 2. Unzip or Place the Skill Files
Extract the downloaded archive and copy the folder containing the SKILL.md file into your skills directory:
```bash unzip my-skill.zip -d ~/.codex/skills/ ```
Or if cloning from GitHub:
```bash git clone https://github.com/username/repo ~/.codex/skills/my-skill ```
3. Verify Skill Metadata
Open the SKILL.md file and ensure it contains proper metadata such as `name`, `description`, and `version`. Codex CLI relies on these fields to load and identify skills correctly. 4. Restart Codex CLI or Start a New Session
The Codex CLI automatically loads any SKILL.md files found in the skills directory at launch. Simply start a new Codex CLI session:
```bash codex ```
The newly installed skills will be available for use. 5. Confirm the Skill Loaded Successfully
Within the Codex CLI, you can list loaded skills by running:
```bash codex list-skills ```
This command outputs all skills currently active. Verify that your new skill appears in this list.
Cross-Compatibility: Claude Code and OpenClaw
One of the strengths of the SKILL.md format is its cross-platform compatibility. SKILL.md files designed for Claude Code or OpenClaw work seamlessly within Codex CLI. This means you can share skills across different AI coding agents without rewriting or adapting the instructions, streamlining your workflow across tools.Practical Tips for Managing Skills
- Organize Skills by Category: Create subdirectories within `~/.codex/skills/` named after skill categories such as `testing`, `deployment`, or `formatting` for easier navigation. - Version Control Your Skills: Keep your custom or downloaded skills under Git version control to track changes and updates. - Regular Updates: Pull updates regularly if you cloned from GitHub to benefit from improvements or bug fixes. - Backup Your Skills Folder: Backup your `~/.codex/skills/` directory to avoid losing customized workflows.Real-World Use Cases
Automating Code Review
Imagine you want Codex CLI to assist in code reviews by running a skill that checks for common anti-patterns or style violations. You can install a skill from Agensi that defines these checks, and Codex CLI will guide you through fixing issues interactively.Continuous Integration Pipelines
Incorporate deployment or build-related skills into your Codex CLI setup. For example, a skill might automate Docker image builds, run tests, and deploy to staging environments, all orchestrated through Codex CLI commands.Custom Developer Onboarding
Create onboarding skills that help new team members understand project structure, coding standards, and setup instructions. These skills serve as interactive documentation accessible directly via Codex CLI.Troubleshooting Common Issues
- Skill Not Loading: Ensure the SKILL.md file is correctly named and placed directly inside a subfolder within the skills directory. - Syntax Errors in SKILL.md: Validate the markdown and metadata formatting to avoid parsing errors. - Conflicts Between Skills: If two skills define conflicting commands or workflows, try isolating them by disabling one temporarily or renaming commands.Conclusion
Installing skills in Codex CLI is a straightforward process that significantly boosts the AI agent’s capabilities. By leveraging SKILL.md files from Agensi, GitHub, or your own projects, you can customize Codex CLI to fit any development workflow, enhance productivity, and ensure cross-compatibility with other AI coding assistants like Claude Code and OpenClaw. Follow the steps and tips outlined above to get started today and unlock the full potential of your AI-powered coding environment.Frequently Asked Questions
What AI tools do developers use most in 2026? Claude Code, GitHub Copilot, and Cursor AI dominate developer workflows. See our Developer Prompts collection.Are AI coding tools worth it? Yes — developers consistently report 20–40% productivity gains on repetitive tasks. Most subscriptions pay for themselves within the first week.
Copilot vs Cursor — which is better? Cursor handles complex multi-file edits better. Copilot integrates into more editors. Most developers who try Cursor prefer it for complex projects.
Is AI-generated code production-ready? AI produces solid starting points, not final code. Always review, test, and refine AI-generated code before deploying to production.
