InterSAGE proposes the trust layer MCP, A2A and ANP left out.

arXiv 2608.13030 points out existing agent protocols specify message exchange but not how an agent proves identity, authorization, advertised capabilities, or accountability after delegation. It adds Persistent Identity, Discovery, Trust Negotiation and Accountability layers via Agent Identity Cards, DID-bound Verifiable Credential manifests, monotonic capability attenuation, and kernel-mediated cryptographic audit trails needing no consensus ledger. Designed to sit alongside MCP rather than replace it, with a comparison against 50+ prior efforts finding none jointly enforcing all four.