Local security audit for AI API relays and LLM proxies: detects prompt injection, model substitution, tool-call rewriting, SSE anomalies, error leakage, and Web3 wallet risks.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).