Posted in Application Security
A November IETF draft says the same thing my framework is built on — and then it goes four steps further…
Schomer — Principal Security Solutions Architect & Founder
A November IETF draft says the same thing my framework is built on — and then it goes four steps further.
Sangam Kumar Das's draft on consequence-bound execution authority (draft-das-consequence-bound-execution-authority-00) puts it in one sentence: permission to propose an operation is not permission to cause its effect.
His architecture holds a proposed operation as a Candidate Act in a Non-Effective State until an independent validator checks its current authority — then a protected Finality Sink re-derives the actual operation and enforces a scoped authorization at the boundary where the effect would become real.
That hold-before-effect shape is what my gates do. The Modbus gate blocks the bad write before it reaches the backend. The shell gate denies before execution. JARVIS pauses state-changing steps for approval. And his five-questions table — identity, tool-invocation, staging, effect-commit, outcome evidence — maps onto those gates almost one for one.
Where the draft goes further than my current gates, honestly:
1. Re-derivation. The Finality Sink re-derives the actual operation from the raw act instead of trusting the proposer's description. My gates check the parsed fields as stated. If a request lies about what it will do, my gate checks the lie.
2. Authority at effect time. The validator checks current authority, including revocation, at the moment of effect. My gates check mostly static rules — the token gate doesn't check revocation at all.
3. Scoped, non-bearer authorization. Bound to one act, one destination, one validity window. My HTTP gate's key is a demo shared secret, and the token gate's tokens are static strings.
4. Outcome evidence from the destination. Receipts from the destination of record, not self-reported logs. My logs are written by the gate itself.
That's not a refutation — it's convergence, and it's a useful one. An independent draft landing on the same architecture tells me the direction is right. The four gaps above are now my upgrade list, in the draft's own terms.
3 reactions · 0 comments