LLM security for AI coding agents -- protect applications from prompt injection, PII leakage, and data exfiltration. Works with Cursor, Claude Code, Codex, Copilot, Windsurf, and any MCP client.
