A multi-agent application-security review harness for Claude Code (and, later, other AI agents). One router skill dispatches to a full offensive-security pipeline that maps a codebase, hunts vulnerabilities with a per-class knowledge base, chains findings into escalations, verifies real impact, and reports to README/JSON/SARIF/doc/PDF.

Detects prompt injection by its effect on a sacrificial canary model, not just pattern matching: untrusted input hits a powerless model first, a behavioral check reads the residue, and it returns block, flag, or pass before your primary model acts. Inbound preflight sensor, not a guarantee.

SAIL V2 (Secure AI Lifecycle) as an agent skill — the full 91-risk catalog for AI/agent gap assessments, security roadmaps, and compliance checklists. Installs on Claude Code, Codex, ChatGPT, Antigravity, and any SKILL.md-compatible agent.

lunar.dev: Agent native MCP Gateway for governance and security