Agent Identity Receipt
A minimal demo showing how an AI action can be identity-bound, policy-checked, audited, and provenance-stamped. The output is an exportable receipt that supports replay and defensibility.
- Agent name and identity used
- Token issuance and scope
- Policy decision in context
- Model call logging
- Audit event emission
- Provenance stamp with fingerprints
The demo is local-first. No vendor integrations required.
Content Lineage Stamp
A minimal prototype showing fingerprinted output, signer identity, and lineage records. It also shows what cannot be proven—which is the point.
- Input fingerprint and output fingerprint
- Signer identity
- Policy reference
- Lineage record ID
- Exportable stamp
Content lineage demo coming soon. The agent receipt demo is available now.
Proof of Concept
These demos prove that identity, policy, audit, and provenance can be instrumented at the action level—not just at the system level.
Reference Implementation
The code is open and minimal. Use it as a starting point for your own instrumentation. Fork it, extend it, adapt it.
Conversation Starter
Show these demos to your security, GRC, and platform teams. Ask: "Can we prove this for our AI systems today?"