🛡️ API Relay Audit - AI Proxy Security Auditor

As developers increasingly rely on AI API relays and LLM proxies, security vulnerabilities in these middle layers are becoming a real concern. This project provides a local security audit tool specifically designed to test these infrastructure components. In short, it tries to expose the hidden risks in API routing. It detects prompt injection attempts, silent model substitutions, malicious tool-call rewriting, SSE anomalies, and potential error leakage. It even includes checks for Web3 wallet risks associated with agent actions. Security is often an afterthought in the rush to deploy AI features. While not as flashy as a new generative model, this tool offers a highly practical defense perspective for anyone managing LLM proxies or gateways in production environments.