What problem does it solve? Manual security reviews of large codebases miss attack classes, duplicate effort, and report unverified findings. This Skill runs a structured, coverage-led vulnerability hunt that assigns attack classes to focused agents, tracks every unit in a ledger, and only reports findings backed by source evidence and bounded local validation. ## Core Features & Use Cases - Coverage-led hunting waves: Splits a codebase into ledger units mapped to attack classes (injection, access control, memory safety, AI/LLM, cloud, client-side, supply chain, and more) so no boundary is silently skipped. - Domain-specific hunting guides: Ships companion references for web/auth, client-side, cloud deployment, data isolation, desktop/mobile IPC, memory safety, and AI/LLM agent targets, each with core disciplines and validation rules. - Strict finding validation: Distinguishes confirmed findings (full source trace plus bounded local proof) from needs_validation candidates, and uses coverage-critic waves to detect gaps before declaring completion. - Use Case: Point it at a multi-tenant SaaS repository; it maps trust boundaries, assigns hunters per subsystem, validates candidates in a sandbox, and produces a deduplicated findings report with remediation guidance. ## Quick Start Ask the agent to run the security-audit skill against this repository and produce a validated findings report.
