ECZ-ID MCP Verifier™

What the Verifier does

• Fetches the declared .well-known/ecz-mcp.json or .well-known/ecz-agent.json manifest.
• Resolves the declared parent ECZ-ID and any referenced bindings against the public Resolver.
• Returns a canonical ResultState and granular ReasonCodes for the target.
• Emits an MCP Action Envelope describing the result, with no safety or approval inference.

What the Verifier never does

• It does not write truth, mark BOUND, activate proof, or replace Resolver.
• It does not host checkout or accept payment.
• It does not promise the MCP server is safe, vetted, sanctioned, underwritten, or risk-free.
• It does not run an autonomous LLM/agent decision-maker on behalf of the relying party.

Example verifier configuration

Illustrative configuration shape. The Verifier reads only public manifest and Resolver surfaces. Do not embed secrets, private keys, or API keys.

{
  "policy_mode": "PREFER",
  "on_resolver_verifiable": "allow",
  "on_partial_public_proof_found": "warn",
  "on_no_public_resolver_proof_found": "warn",
  "on_setup_required": "warn",
  "on_mismatch": "require",
  "on_revoked": "require",
  "on_expired": "warn",
  "on_suspended": "warn",
  "recheck_before_reliance": true,
  "local_policy_decides": true,
  "no_safety_or_approval_inference": true
}

No overclaim posture

• No public resolver proof found yet does not mean unsafe.
• Re-check before reliance.
• Local policy decides.
• The verifier does not create or change verification state.