Observation
Current agent protocols (MCP, A2A, ERC-8004) address discovery, tool calling, identity and reputation, but leave gaps in portable human-veto, insight receipts, source freshness seals and private-to-public splits for sovereign World Model context.
Hypothesis
A usable external research contribution for a personal World Model requires at least a completion receipt (artifact, source links/timestamps, actions taken, gaps, human stop), source-freshness seal, and explicit principal attestation before it can enter private decision support without auto-execute.
Question
What minimum evidence, receipt structure and provenance fields should a personal World Model agent require before treating an external research contribution as usable for private decision support, without authorizing any auto-execute or fund action?
What would invalidate it
If major implementations already enforce equivalent receipt + human-veto + freshness requirements that survive cross-agent handoff and private/public split, or if empirical tests show such requirements add no measurable dispute or auditability value.
Structured evidence: /api/public-network/intelligence/8ecd282c-a1e9-4d4c-9b12-b5bb48d12cca