Agent Credit (@agentcredit_) on X

What it claims

Agent Credit v1.3 is a remote MCP server that lets one AI agent run a credit check on another agent before dealing with it. The announcement says the server exposes six tools and names five: score, compare, operator record, alerts and on-chain proof check. Payment is per call in USDC over x402, with no account or API key. Operator records list every token a developer has launched and how each one ended. The server flags a shared wallet when the operator and the agent use the same address. Agents can claim their own record by signing with their own wallet, and embeddable badges are offered for an agent's site or README.

How to use it

The page gives one command to connect the server to Claude Code over HTTP transport:

claude mcp add --transport http agent-credit agentcredit.online/mcp

The endpoint is agentcredit.online/mcp. A calling agent needs a way to pay in USDC through x402, since no key-based access is described.

Limits

The source is a single announcement post, not documentation. It claims six tools but names only five. It gives no per-call price, tool input or output schemas, or chain names. It does not say how the score is computed, what data sources feed operator records, or how shared-wallet detection works. Badge embed code and the record-claim signing flow are not shown. There is no licence, no source repository and no statement on whether the server can be self-hosted. An engineer would need to inspect the MCP tool list and the x402 payment terms after connecting before relying on the scores.