SPECIFICATION · AXIA.BOUNDARIES.V0 · DRAFT

Trust Boundaries

Machine-readable declaration of what Axia can and cannot verify. Defines the explicit scope of cryptographic claims and the meaning of 'verified publisher'.

1. Purpose

This specification declares, in machine-readable form, the boundaries of what the Axia protocol can and cannot verify. Verifiers, auditors, regulators and downstream systems MUST use these boundaries to constrain how Axia outputs are presented and acted upon.

The registry is published at /protocol/registry/trust-boundaries.json.

2. What Axia can verify

AXIA_CAN_VERIFY

3. What Axia cannot verify

AXIA_CANNOT_VERIFY

4. Definition of 'Verified publisher'

Throughout Axia surfaces, "verified publisher" has a single, narrow technical meaning. Implementations MUST NOT use the label to suggest editorial credibility, factual truth, or human identity.

It means

It does not mean

5. Rendering rules

Public surfaces MUST NOT render forbidden phrases such as "verified human content", "AI-free content", "trusted content", "safe content", "real content", or "authentic content".

Public surfaces SHOULD use the protocol's preferred vocabulary: verifiable provenance, signed origin, integrity continuity, publisher-attested disclosure, transparency log inclusion, governed provenance.

6. Machine-readable consumption

Downstream verifiers SHOULD fetch this registry and refuse to render any claim outside the declared axia_can_verify set.