toby-bridgestoby-bridgesCommunity·1 Agent Skills Included

api-relay-audit

Audit third-party AI API relays for hidden tampering locally

Audits third-party AI API relays and LLM proxies for hidden prompt injection, model substitution, context truncation, and tool-call rewriting. Runs 14 local checks with only Python and curl, so your API key never leaves your machine or touches a third-party web tool. Generates a structured Markdown report with a clear LOW, MEDIUM, or HIGH risk verdict before you trust a relay with production or wallet traffic.
npx skills add toby-bridges/api-relay-audit --all -g -y

All Skills in This Repository (1)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install API Relay Audit?

Run `npx skills add toby-bridges/api-relay-audit --all -g -y` in your terminal to install the skill globally for your AI agent.

How to check if an AI API relay is safe?

Run the audit with your relay key and base URL, and it produces a Markdown report rating the relay LOW, MEDIUM, or HIGH risk across 14 tamper checks.

What does API Relay Audit detect?

It detects hidden prompt injection, prompt extraction, identity and model substitution, context truncation, tool-call rewriting, error leakage, and stream anomalies.

Does it work with OpenAI and Claude compatible proxies?

Yes. The script auto-detects Anthropic and OpenAI API formats and adapts its probes without any manual configuration.

Is my API key safe during the audit?

Yes. The audit runs fully on your machine and sends your key only to the relay URL you choose, never to any third-party service.

Related Repositories in Software Engineering

View All in Software Engineering