agentsclimarketplace
Sign in

Codex agent dev kit

Skill yigityildiz0/universal-ai-skill-library/skills/platforms/codex/codex-agent-dev-kit

Convert Claude Code agent-kit setups into Codex-compatible skills, project instructions, guardrail scripts, command workflows, subagent guidance.From its SKILL.md

Install
npx -y skills add yigityildiz0/universal-ai-skill-library --skill codex-agent-dev-kit

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.5 KB, 739 tokens by cl100k_base, as published. Nobody here has run it

Codex Agent Dev Kit

Use this skill to translate Claude-oriented agent setup material into Codex-native files. Do not copy .claude files verbatim unless the user explicitly wants archival copies; map each concept to a Codex-supported artifact.

Default Targets

  • Global reusable skill: %USERPROFILE%\.codex\skills\<skill-name>\SKILL.md
  • Project memory and conventions: <repo>\AGENTS.md
  • Repo-local plugin bundle: <repo>\plugins\<plugin-name>\.codex-plugin\plugin.json
  • Plugin marketplace entry: <repo>\.agents\plugins\marketplace.json
  • Deterministic guardrails: helper scripts under a skill or plugin scripts/ folder

Migration Workflow

  1. Inventory the source material.
  • Identify Claude concepts: CLAUDE.md, .claude/commands, .claude/hooks, .claude/rules, .claude/agents, .claude/settings.json, plugin manifests, and screenshots.
  • Separate reusable behavior from repo-specific conventions.
  1. Choose the Codex destination.
  • Put reusable workflows in a global Codex skill.
  • Put repo architecture, commands, test expectations, naming, and local conventions in AGENTS.md.
  • Put shareable bundles in a local Codex plugin only when the user needs plugin UI discovery or team distribution.
  • Put ongoing skill maintenance behavior in skill-updater rather than mixing health checks into every skill.
  1. Translate, do not clone.
  • CLAUDE.md memory becomes AGENTS.md or skill instructions.
  • Claude slash commands become named workflow sections, default prompts, or executable scripts.
  • Claude hooks become explicit guardrail scripts and preflight checks; a skill cannot enforce every tool call by itself.
  • Claude rules become AGENTS.md project rules or skill body guidance.
  • Claude agents/subagents become delegation guidance. Codex subagents are used only when the user explicitly asks for delegation or parallel agent work.
  • Claude settings.json permissions and model values do not port directly. Do not weaken Codex safety settings during migration.
  • Claude plugins become Codex plugins with .codex-plugin/plugin.json, skills, optional assets/scripts, and a marketplace entry.
  1. Validate the result.
  • Run the Codex skill validator for every skill folder.
  • Validate plugin JSON, marketplace JSON, and relative paths.
  • Run bundled guardrail scripts with a harmless target before relying on them.

Reference Files

  • Read references/claude-to-codex-map.md when converting screenshots or .claude folders.
  • Read references/codex-plugin-layout.md when creating or fixing a Codex local plugin package.
  • Use $skill-updater after migration when a generated skill fails, becomes stale, references old models, or needs a health audit.

Guardrails

  • Never import Claude permission allowlists as Codex permissions without reviewing each entry.
  • Do not add destructive commands, force-push workflows, or hidden automatic commits.
  • Prefer small, explicit scripts over broad "run everything" automation.
  • Keep skill frontmatter to name and description; Codex uses those fields for triggering.
  • Keep agents/openai.yaml UI metadata short and synchronized with SKILL.md.

Verification Helper

Use scripts/check_codex_agent_kit.ps1 to check that a skill or plugin bundle has the expected Codex files.

Changelog

  • [2026-05-08] [HEALTH] Added skill-updater routing so migrated skills can be maintained through failure-driven and usage-based updates.

What ships with it: 4 files

6.5 KB alongside SKILL.md, 1 of them executable

agents/

references/

scripts/

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.