VIAS-1.0 · Protocol Overview
An evidentiary independence protocol for autonomous agent execution.
VIAS-1.0 defines how execution records submitted by autonomous agents are sealed outside the agent runtime and made independently verifiable. VIAS attests to the records submitted to it — their integrity, provenance, and stated coverage — not to whether the agent's action was correct or permitted.
Every record checked against all six.
Every sealed record carries the same six dimensions of evidence.
- Input
- What the agent was given to act on
- Origin
- Which model and version acted, and where the record came from
- Response
- What the agent produced
- Handoff
- How the action links to other agents in a chain
- Scope
- What the agent was authorised to access, and what access was reported
- Authorisation
- Who approved the action, and under which policy
Verifiable without VIAS.
VIAS is designed so that any party in possession of an evidence record can recompute the record's hash from its canonical fields and confirm it matches the sealed value. Verification does not require access to the VIAS platform, the agent vendor, or any third-party service. The signed-export verification chain is pending end-to-end testing; until it completes, offline verification should be treated as designed-for, not confirmed.
Full specification available under NDA.