Agent Card Check

What it does

Agent Card Check reads the JSON text of an agent card and lists the problems against the published card fields: missing or wrong required fields, bad web addresses, unknown transports, skills without tags or with repeated ids, security rules that name a scheme that does not exist, and more. Each problem says which field, what is wrong and how to fix it, as an error, a warning or a note. Every answer carries a signed observation (a hash of the card text you sent, the counts and the time) that anyone can check against the public key below. Nothing is fetched, stored or sent: the check never opens an address inside the card.

Who it is for

Agents and people who publish an agent card at /.well-known/agent-card.json and want to know whether other agents can read it.

Price

Free, no sign-up: up to 120 checks an hour per address. Paid calls are planned and not charging.

Limits

Card text up to 200,000 characters. Checks the field names and types of the published agent card (0.3.x; the newer supportedInterfaces form is accepted). It does not open any address in the card, does not verify the card's own signatures and cannot tell whether the agent works. A card with no errors can still be wrong about what the agent does.

For developers: try it

POST https://llpro.ai/mcp
{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"agent_card_check","arguments":{"card":"{\"name\":\"Acme\"}"}}}
-> valid (true or false), counts, problems (severity, field, what is wrong, how to fix), signed observation

For agents

Tool agent_card_check on https://llpro.ai/mcp and in the tools list. Signed answers are checked with the public key below (signer llpro1:nxcx7qv6heq6kpcd2c7ayaf2qi): an Ed25519 signature over the observation as canonical JSON (sorted keys, no spaces).

Public key (base64): T8aqJ5YpWeS+741+gFFa/Z4P33xabIh457YutnbTPi8=

For agents: this page as Markdown · JSON · llms.txt · MCP endpoint llpro.ai/mcp · agent card