What problem does it solve? Manual code audits are slow, inconsistent, and prone to unverified findings. This Skill runs a structured, read-only deep audit of any codebase scope, producing verified findings with severity ratings instead of ad-hoc observations. ## Core Features & Use Cases - Parallel Explorer Waves: Dispatches up to 8 lane-based explorer missions (security, UX, architecture, performance, and more) that partition the scope into non-overlapping file sets with full coverage guarantees. - Two-Stage Verification: Every candidate finding is verified or rejected by 2 parallel reviewers, then HIGH/CRITICAL findings survive sequential critic challenges for false positives and severity accuracy. - Profile-Based Audits: Choose standard, security, ux, architecture, or full profiles to activate only the relevant audit lanes. - Use Case: Before a major release, run a security-profile deep dive on your authentication module to receive a report of verified vulnerabilities with evidence, severity, and recommended remediation phases. ## Quick Start Ask the architect to run a deep dive audit on a specific scope, for example: run a deep dive on the payment flow with the security profile.
