Research · August 24, 2026
Human-in-the-Role: a technical note on binding authority
How role bindings turn non-repudiation from a procedure into a structural property of execution.
Summary
A short companion to Vol. I setting out how the Role Identity Fabric issues, scopes, and expires the authority an agent exercises, and what an evidence package produced from those bindings actually contains.
Full text
The record
The claim
Non-repudiation in agentic systems is usually procedural: a policy says a human approved the action, and a log is offered as proof. Procedure is contestable. Structure is not.
Binding, scope, expiry
A role binding names the professional, the scope of authority granted, the systems in reach, and the moment the binding lapses. An agent operating outside that envelope does not get refused at review time; it cannot act at all.
What the evidence package contains
For each Verified Execution Cycle: the directive as issued, the role binding in force, the reconciliation the Logic Scrubber performed, the data boundary enforced, and a hash chained into the Immutable Audit Ledger.
Corrections welcome, in writing.
If a figure we published is wrong, we want the citation. Research correspondence goes straight to the founding team.
