Tyche Institute · attestable agency

From presence to authority

Four rungs of machine-checkable legitimacy. Read left to right: each rung anchors a stronger claim than the last, and the question worth defending is climbing.

presence authority
Mechanism
The question it answers
Anchored by
How it breaks
BotGuardbot detection
Is there a real human or genuine browser?
environment heuristics — no anchor
solve it / reverse-engineer it
PATprivate access token
Is this a genuine client device?
hardware (Secure Enclave)
attests the client, not the action
PACTagent legitimacy
Is this a legitimate agent?
no hardware anchor
forge / replay the claim
AEPaction evidence
What did it do, under whose authority, in mandate?
issuer key + TPM + scope
the question that remains
The attack surface migrates: presence→authority — once agents act on their own, "is there a human?" stops being the question.

Why the ground is shifting

The first rung is a moat that large language models are draining. The once-esoteric TLS fingerprinting that made bot-detection an expert's game — cipher ordering, JA3/JA4, automation signals — is now a couple of prompts away. Hardware tokens raise the floor to is the device genuine; agent-legitimacy schemes ask is the agent legitimate, without a hardware anchor. But an agent acting on your behalf raises a different question that none of the first three answer: what did it do, under whose authority, and was it in mandate?

That is the rung Tyche's work sits on. A per-action evidence package, bound to a scoped, signed mandate, lets an offline third party adjudicate a single action — and reject one that is correctly signed but out of mandate. The live demonstrator: Who Said the Agent Could Do That?

Diagram by Tyche Institute; part of the Lab. Self-contained, no external calls. See also: the EU digital acquis as one graph.