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.
dsh plugin --profile web add dsh-api-relay-audit
View the source on GitHub