Personal, reusable skills for AI coding assistants (Claude Code / Codex).
Each folder is a self-contained skill: a SKILL.md plus, where useful, a
references/ deep-dive rule sheet and an agents/openai.yaml for Codex.
| Skill | What it does |
|---|---|
python-code-format |
Python 3.12 style decisions — formatting, naming, docstrings, logging, resource handling. Includes a detailed references/ rule sheet. |
cpp-code-format |
C++/C (C++20) style — layout, naming, headers/includes, comments, error handling, logging messages. Includes a detailed references/ rule sheet. |
prompt-engineering |
Writing and optimizing prompts, commands, hooks, and sub-agent instructions for LLM interactions. |
create-hook |
Author a hook (.json) to enforce policy or automate agent lifecycle events. |
incremental-change-cycle |
Run an incremental change cycle — checklist continuity and repeated review checkpoints — for long-running, high-token work. |
Drop a skill folder into your assistant's skills directory and invoke it by name:
- Claude Code:
~/.claude/skills/<skill>/ - Codex:
~/.codex/skills/<skill>/
MIT © 2026 WENHUIZ