The authority gap
The authority boundary
On one side, a proposal. On the other, a real state change. DETERMA sits in the middle and revalidates the exact action at execution time — external to the agent and external to the model.
01
Proposal
02
DETERMA — Execution Authority
03
State change / external effect
Governed lifecycle
AI may propose. Authority decides. Constrained executors mutate. Evidence proves.
Evidence and validation are not authority.
01
Proposal
02
Validation
03
WAITING_APPROVAL
04
Human approval context
05
Exact scope binding
06
State Witness where required
07
Fresh single-use Execution Release
08
Constrained Executor
09
Append-only audit
10
Post-execution verification
What DETERMA is, and is not
- DETERMA is execution-authority infrastructure.
- It is not a generic agent framework.
- It is not a direct-AI mutation system.
- It is external to the agent and to the model it evaluates.
Authority invariants
- Proposal and execution are separated.
- Agents and workers hold no general mutation authority.
- The orchestrator is not the executor.
- WAITING_APPROVAL is a hard stop.
- Human approval alone is insufficient.
- An approval cannot be replayed after consumption.
- Ambiguity, drift, replay, missing scope or audit failure block.
Failure model
Approval drift
The requested action is no longer the action that was approved.
Stale head / base metadata
The state the approval was bound to has moved.
Unresolved findings
Open findings contradict the requested effect.
Scope drift
Requested parameters exceed the reviewed bounds.
Replay
A consumed single-use release is presented again.
Missing readback
The effect cannot be confirmed after execution.
Missing post-change verification
No verified comparison of intended and actual effect.
Why existing controls are not enough
Each layer answers a different question.
Identity proves who.
DETERMA validates what may happen now.
Sandboxing controls where.
DETERMA controls authority.
Orchestration keeps work moving.
DETERMA decides whether it may continue.
Observability explains what happened.
DETERMA can block before the side effect.
Next step
Which action would you want AI to be able to perform safely?
Tell us the workflow and the action. We will tell you whether a bounded evaluation makes sense — and what evidence it would produce.