Principles
Arguments, not slogans.
Each principle below is a position we are prepared to defend in detail, and each one is enforced somewhere in the architecture rather than asserted in copy.
- Independent verification
- Least authority
- Evidence over claims
Principle 01
Independent verification
Work is not done because the component that did it says so. It is done when a party outside the execution path attests that the observable state matches the claim.
A claim of completion is evidence about the claimant, not about the system. Verification only means something when the verifier is structurally incapable of having caused the outcome it judges.
Read in fullPrinciple 02
Least authority
Authority is granted narrowly, explicitly and in advance. Anything outside the declared envelope is refused, not negotiated.
Every component receives the narrowest authority that lets it do its job, declared before it starts and enforced at admission rather than argued about at runtime.
Read in fullPrinciple 03
Evidence over claims
Every consequential outcome is bound to an artefact that can be re-examined independently: a commit, a prepared-state digest, a durable transition or a signed receipt.
Prose summaries decay, get rewritten and cannot be re-checked. Evidence is bound to an exact commit or prepared state and can be inspected long after the session that produced it has ended.
Read in full
What they add up to
One system, defined precisely
- Definition — Proof & State
- Proof & State is AI work accountability infrastructure: DoneState executes under declared authority, AgentProof authorises consequential actions and signs receipts, and OpsTruth verifies the result read-only from outside the execution path.
Give agents authority. Keep the proof.
If you disagree with a principle, the argument is on the page — take it apart.