Local security audit tool for AI API relays, LLM proxies, and Web3 agent pipelines.
Detect prompt injection, model substitution, and Web3 wallet risks before sending production traffic.
API Relay Audit is a local security scanner that evaluates third-party AI API relays, mirrors, and gateways for hidden prompt injection, context truncation, and model substitution. You should use it if you route production or agentic traffic through a third-party Claude-compatible or OpenAI-compatible proxy and need to verify its integrity before deployment.
The tool executes structured prompt and response checks using a standalone Python script or a DeepSeek Harness plugin, generating a local Markdown report with evidence for every test step. It tests instruction override, package-command alterations, error-response leakage, and SSE stream anomalies without sending your API keys anywhere except your chosen relay.
It suits AI engineers, security auditors, and developers evaluating third-party LLM gateways or Web3 agent relay endpoints. It does not suit teams looking for a fully managed cloud safety certification or automated real-time operational monitoring.
Run a zero-dependency Python script that sends probes exclusively to your chosen relay endpoint.
Check relay behavior regarding ETH transfer guidance and signed-transaction refusals via a dedicated profile.
Integrate the audit suite directly into DSH or TUI workflows using existing provider credentials.
Generate a reviewable Markdown report with per-step evidence and low, medium, or high severity summaries.
The tool is free under its open-source license, and the only cost of running it is the consumption of metered tokens from your chosen API provider during audit executions.
896 open remote AI-training roles, updated November 16, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
These reviewed AI tools do the same job instantly — from our scored directory.
No, your API key is sent only to the specific relay URL you pass into the script.
Yes, you can use the web3 or full profile to check wallet-sensitive relay behaviors and transfer guidance.
No, the tool produces a local evidence report and explicitly does not claim to certify relay safety.
See Zekai first in Google