Trust infrastructure for agentic AI — at the edge.
Fingerprint agent traffic, verify signed behavioural attestations, score trust, and enforce allow/deny in real time. The shared security layer behind Whimsy and Codero.
Role
Agent trust layer
Year
2026
Surface
Edge service · SDK
Status
Internal — powering Whimsy and Codero
01The problem
Agentic systems need trust, not just authentication.
When the actor on the other side of an API is an autonomous agent, the question is no longer "is this token valid?" but "is this behaviour consistent with what this agent claimed to do?".
VeriLink fingerprints agent traffic, verifies signed behavioural attestations, calculates a trust score, and enforces allow/deny decisions in real time — before the request reaches the protected service.
One key. A stack of cards. A fingerprint. That's the model.
02The approach
Decide at the edge, observe everywhere.
VeriLink sits between agent traffic and the services that matter. Decisions happen in microseconds; the receipts are durable and inspectable.
It is the same layer Whimsy uses to gate cross-cloud actions and Codero uses to scope what an AI reviewer is allowed to touch.
Specifications
“Identity is who you are. Trust is what you are about to do.”
— Studio note
Previous work
← SlopgateNext work
pipguard-cli →