Connector design question: is a verifiable-attestation identifier a first-class citation source for ai-governance-legal?

Building an MCP connector for the ai-governance-legal plugin and want to confirm the intended design before opening a PR against .mcp.json.

CONNECTORS.md asks for "provenance in results — the source, date retrieved, and a citation-ready identifier" so the plugins can tag every cite by source (and otherwise flag [verify]). Our connector is read-heavy (list_agent_inventory, fetch_permission_map, get_audit_attestation) and returns live AI-agent governance telemetry where each finding carries a cryptographically verifiable attestation identifier (an on-chain proof URL + hash) rather than only a plain citation string.

Two questions:

1. Does the [verify] / source-tagging path treat a verifiable-attestation identifier as a first-class citation source, the same way a registered research DB cite is treated — or is source-tagging reserved for a known allowlist of research connectors (CourtListener, Trellis, Descrybe, etc.)?

2. Beyond the retrieval-quality + injection-resistance checks, are there requirements for a connector whose differentiator is verifiable provenance of the returned governance data (e.g., a way to surface the verify-URL into the deliverable, or expectations about attestation formats)?

Happy to share the tool schema and a worked example. Thanks for open-sourcing this — the connector contract is clear and the read-heavy + provenance design maps well to what we already produce.