CI-native security testing for MCP servers. Attack simulation, schema drift detection, and health scoring before agents depend on them.

Local-first MCP security scanner for AI-generated apps. Scan → fix → rescan from Claude Code, Cursor, Codex, and other agents.