Blog
The KYE Protocol™ blog: doctrine, plainly argued.
Series: ― Is Not Authority. Eight adjacencies, one boundary; every post resolves to the same place: Authority Finality™. Each post reads in under two minutes.
Decision Provenance Is Not Authority Provenance
Tracing how an AI outcome was produced is not proving the actor was entitled to cause the effect. Where decision provenance stops and Authority Finality™ begins.
Read23 July 2026The Agent on Trial
A successful AI-agent action is not a legitimate one. Put the action on the stand and ask the one charge a dashboard never answers — was it authorised? The verdict is Authority Finality™, dramatized in the Authority Court.
Read22 July 2026Decision-Admissibility Is Not Action-Authority
Compiling a clause into code proves the rule is satisfiable — not that anyone may act on it. Where policy-as-code stops and Authority Finality™ begins.
Read21 July 2026Reasoning Control Is Not Authority
Shaping how a model reasons is not deciding whether its conclusion may become an action. Where inference-time control stops and Authority Finality™ begins.
Read20 July 2026An API Key Is Not Authority
A key proves the caller can reach the API; it never proves the caller should have made the call. Where credentials stop and Authority Finality™ begins.
Read19 July 2026Identity Is Not Authority
Knowing who an AI agent is answers the door; it does not decide what the agent may do. Where agent identity stops and Authority Finality™ begins.
Read18 July 2026A Permit Is Not Authority
A permit says the action is allowed. Authority says this principal may take it. Where runtime guardrails stop and Authority Finality™ begins.
Read17 July 2026A Trace Is Not Authority
A trace tells you what happened. Authority decides whether it was permitted to. Where agent traces stop and Authority Finality™ begins.
Read16 July 2026Observability Is Not Authority
Watching an AI agent act is not the same as deciding it was allowed to. Where observability stops and Authority Finality™ begins.
ReadMoreResearch in the library, product updates in the changelog and the updates feed.
- A Permit Is Not AuthorityA permit says the action is allowed. Authority says this principal may take it. Where runtime guardrails stop and Authority Finality™ begins.
- A Trace Is Not AuthorityA trace tells you what happened. Authority decides whether it was permitted to. Where agent traces stop and Authority Finality™ begins.
- An API Key Is Not AuthorityA key proves the caller can reach the API; it never proves the caller should have made the call. Where credentials stop and Authority Finality™ begins.
- Decision Provenance Is Not Authority ProvenanceTracing how an AI outcome was produced is not proving the actor was entitled to cause the effect. Where decision provenance stops and Authority Finality™…
- Decision-Admissibility Is Not Action-AuthorityCompiling a clause into code proves the rule is satisfiable — not that anyone may act on it. Where policy-as-code stops and Authority Finality™ begins.
- Identity Is Not AuthorityKnowing who an AI agent is answers the door; it does not decide what the agent may do. Where agent identity stops and Authority Finality™ begins.
- Observability Is Not AuthorityWatching an AI agent act is not the same as deciding it was allowed to. Where observability stops and Authority Finality™ begins.
- Reasoning Control Is Not AuthorityShaping how a model reasons is not deciding whether its conclusion may become an action. Where inference-time control stops and Authority Finality™ begins.
- The Agent on TrialA successful AI-agent action is not a legitimate one. Put the action on the stand and ask the one charge a dashboard never answers — was it authorised? The…