Deep Audit

Post-execution audit and auto-heal. Spawns specialist sub-agents to find issues, then fixes them. Runs after implementation is complete — never during.

When This Skill Activates

Trigger after execution completes:

• User says: "done", "finished", "all done", "ready to review", "ship it"
• User explicitly requests: "audit", "code review", "evaluate", "deep audit"
• Another skill's closing checklist runs (e.g., TDD closing audit)

Do NOT activate mid-implementation. This skill is a closing gate, not a blocker.

Phase 1: Context Gathering

Before spawning agents, establish scope:

1. Run git diff --name-only HEAD to get the list of changed files. If the working