DEMO — SENTRYFOX — REGULATED COMMERCE CONTROLRepresentative data. Nothing here is a real merchant or a real transaction.ALL DEMOS →
SENTRYFOX
EVIDENCE
Decision Replay
Re-run a past decision against the policy version that was in force at the time.
Replay DEC-4C2A-91F7-0B88-E31D
Original outcomeBLOCKED
Policy version at decisionv18
Policy version nowv18 — unchanged
InputsAs captured, from the evidence record
Replay outcomeBLOCKED — identical
Rule firedBUY-07
Why replay against the old version
- A decision is bound to its policy
- Replaying under today's rules answers a different question: what would we do now. Replaying under the version in force answers the one that gets asked — was this correct at the time.
- The engine is deterministic
- Same inputs, same policy, same outcome, every time. A replay that could differ would make the original unexaminable.
- The reference is not
- Decision references are sixty-four random bits. They identify a decision; they never encode one.
Replay is read-only in the strongest sense
It produces no new decision, changes no state and writes nothing except the fact that a replay was run. A replay that could alter an outcome would be a re-decision wearing the word replay.
DEMO — FIXTURES ONLY, NO DATABASE