selfaudit
The selfaudit skill pauses work to examine recent output through a structured first-principles review, categorizing findings as evidence-based facts, assumptions, potential errors, and highest-risk bugs, then recommending patches or confirming stability. Use this when completing significant work segments to validate logic, catch overlooked issues, and prioritize remediation before proceeding.
git clone --depth 1 https://github.com/sterlingcrispin/claude_slash_commands /tmp/selfaudit && cp -r /tmp/selfaudit/selfaudit ~/.claude/skills/selfauditSKILL.md
Pause and do a first-principles audit of recent work. List: 1) What is true (supported by evidence) 2) What is assumption 3) What could be wrong 4) Highest-risk bug 5) Next highest priority task and why Then either patch issues or state clearly why no changes are needed.
Send recent work to OpenAI Codex CLI for an independent audit/review
Send specific files to OpenAI Codex CLI for an independent audit/review
Deep audit producing a handoff document to get another agent fully up to speed
Concise checkpoint on hypothesis, changes, metrics, and next action
Pair programming continuation
Commit and push changes with auto-generated summary
Start a ralph style looping session
Work on a single highest-priority feature from current_ documents, test it, update docs, and commit