ngreenwall

ngreenwall/claude-skills

Instructions file

Instructions for ngreenwall/claude-skills, covering claude-skills, what this is and key decisions.

ngreenwall/claude-skills

Skill Claude CodeCodex

Manual context-drift audit. Confirms your global context file (the standing instructions loaded on every session) is still loaded and being followed, and flags where context dropped. Run ONLY when the user explicitly types /drift-check or says "drift check," "context check," "are you still following my context," or…

ngreenwall/claude-skills

Skill Claude CodeCodex

End-of-session workflow to leave durable project context for the next chat. Use whenever the user says "handoff," "wrap up," "end of session," "before we end," "close out this session," "let's wrap," "we're done," "wrap this up," "closing out," "starting a new session," or "before I start fresh." Do NOT use for…

ngreenwall/claude-skills

Skill Claude CodeCodex

Audits skills (SKILL.md files), your global context file, and CLAUDE.md/AGENTS.md files against Anthropic's current Agent Skills guidance and the behavior of current models (e.g. Sonnet 5, Opus 5). Use when the user says "audit for new models," "audit my skills against latest Anthropic guidance," "are my skills still…

ngreenwall/claude-skills

Skill Claude CodeCodex

One-time project bootstrap workflow for new repos. Use whenever the user says "initialize project," "set up this repo for Cursor and Claude," "new Cursor project," "set up Cursor," "bootstrap docs/context," or "new project setup." Prefer this before first handoff in a repo. Do NOT use for end-of-session updates in…

ngreenwall/claude-skills

Skill Claude CodeCodex

Static-analysis QA pass over instruction/config files (SKILL.md, CLAUDE.md, your global context file, command files), not code. Spawns an Opus subagent per file to check for contradictions, dead logic, unstated dependencies, and style violations, then proposes fixes for confirmation before applying. Use when the user…

ngreenwall/claude-skills

Skill Claude CodeCodex

Audits instruction/config files (SKILL.md, CLAUDE.md, your global context file, slash command files) and this repo's skill architecture for token bloat, then proposes fixes: rewrite, split, merge, add navigation, or remove. Use when the user says "audit tokens," "check token bloat," "shrink this file," "audit the…