Probe

About Probe

Probe is a compliance scanner for AI agent APIs and MCP servers. You paste a URL and it runs forty-four checks in about thirty seconds, returning a 0-100 score, a letter grade, and a README badge, with no account required. The check set is unusually specific to the agent ecosystem rather than being a generic web scanner with AI branding. It covers the x402 agent payment standard by auto-discovering /.well-known/x402.json and validating payment metadata; MCP endpoint discovery, tool declarations, and transport security; Google's A2A protocol via agent.json discovery and skill declarations; ERC-8004 onchain agent identity registration on Base and Ethereum; and llms.txt and AI plugin discovery. Alongside those it runs conventional infrastructure checks — HSTS, CSP, X-Content-Type-Options, X-Frame-Options, Referrer-Policy, CORS, rate limiting, DNSSEC, CAA, DMARC and SPF, TLS, OpenAPI spec validation, API versioning, and Content-Type consistency — plus a regulatory cluster covering EU AI Act human oversight under Article 14, AI disclosure and risk classification, GDPR privacy and data-processing disclosure, FATF Travel Rule VASP disclosure, and ten voice-AI-specific checks for synthetic voice labeling, call recording consent, and operator identity. The grading is framed around listing readiness: A means listing-ready with minor fixes, C means the endpoint will fail x402 validation, D and F mean critical gaps. Subscribing adds up to 24 scans a day with email alerts. The vendor states the product is entirely free with no paid tier, funded by the compliance dataset it builds.

Key Features

Tags

More AI Agent Infrastructure tools