Test the contract.
Then test the agent.
A technical snapshot for an existing A2A runtime: public discovery, AgentCard validity, a real negotiated message call, identity, signatures, security, and — when Agenstry already monitors it — reliability history.
Canonical and observed AgentCard locations.
A2A schema, v1 interfaces, skills, and security.
Negotiated JSON-RPC message/send response.
JWS, provider identity, latency, and history.
A readiness check, not a readiness theatre.
The snapshot only scores controls the scan can actually observe. Unsigned cards and absent historical uptime remain visible, but “not measured” is not silently treated as failure. Repeated monitoring adds the temporal layer: availability, drift, stable latency, and identity verification. Read the public methodology →