Skip to content
Concepts: all pages

Concepts

Reciprocal trust

Reciprocal trust means both sides of a machine interaction can resolve each other before relying on each other: the agent checks the tool, and the tool checks the agent.

Accountability Chain:IdentityCurrent stateRe-verify

Both sides resolve

Most trust checks run one way: a client checks a server. In the machine economy both sides act. An MCP server calling back into an agent, an API receiving an agent’s request, two agents negotiating: each side has its own reason to know who is on the other end, and its own policy for what it requires.

The Reciprocal Reliance Envelope is the published, read-only shape for writing down what each side found when it resolved the other. Each half reads PUBLIC_RECORD_FOUND, NO_PUBLIC_ECZ_ID_FOUND or RESOLVER_UNREADABLE, and the third is not the second. Its external_authorisation field is pinned at not_determined_by_eczid.

Reciprocal receipts

Core’s Receipt contract defines a reciprocal trust receipt type: a receipt that captures both sides’ relevant trust context. It is a receipt capability, not a separate product, and it does not imply mutual legal acceptance or co-signature unless those artifacts exist. Like all receipts, it is not publicly retrievable today.

This page is explanation, not authority. Current state is only ever the live Resolver record. Live facts on this page were measured read-only on 2026-10-10.