Personal, cross-project Codex skills. This repository is the source of truth; installed copies are generated from it.
From the repository root, run:
python3 scripts/install_skills.pyThe installer copies skills to ${CODEX_HOME:-$HOME/.codex}/skills. Set CODEX_SKILLS_DIR to override the destination. Run it after cloning, pulling, or changing a skill. It replaces only same-named skill folders and leaves other installed skills alone. Deleting a skill from this repository does not uninstall its local copy.
These skills are maintained in this repository and installed by the script above.
| Skill | Useful for |
|---|---|
| performance-aware-coding | Spotting likely runtime costs during implementation or review, then validating meaningful optimizations with evidence. |
| brand-domain-research | Brainstorming brand and product names, suggesting matching domains, and checking registration records with RDAP when requested. |
Add skills/<name>/SKILL.md with the required name and description front matter. Include agents/openai.yaml or supporting files only when needed. Keep changes evidence-based and review them before installing.
These skills are maintained outside this repository and are not copied or installed here. Check each upstream source for current installation and license details.
| Skill | Useful for | Upstream source |
|---|---|---|
browse-x |
Reading public X posts, threads, profiles, and search results as Markdown or JSON. | pc-style/x-md |
security-best-practices |
Applying language- and framework-specific secure coding guidance. | openai/skills |
security-threat-model |
Mapping a repository's trust boundaries, assets, attacker paths, and mitigations. | openai/skills |
prune-tests |
Finding brittle or tautological tests and removing them when cleanup is authorized. | riffingonsoftware/agent-skills |
security-audit |
Running a structured, source-grounded security audit. | cloudflare/security-audit-skill |
This repository is licensed under the MIT License. See LICENSE.