✔ Passed
- Case
journeys/journey-spend-control-ai-costjourney- Run at
- 2026-11-05 00:42 UTC
- Ran on
- our own install (dev)
- Checks fingerprint
- 021c1cfa4ac0
- Evidence id
- ev_mw2tmmpc_2a98f4
proof-engine run mw2tmm
Journeys: Spend control AI cost
A machine ran this test against a real Quox install and recorded the result. Nobody marked it by hand.
Passed
What this test checks
- no_toolQuox did not use a tool that was ruled out
- not_saysThe answer makes none of the ruled-out claims
- says_anyThe answer says one of the accepted things
- toolQuox used the expected tool, and it succeeded
The fingerprint is a hash of the test's checks: if the test ever changes, so does the fingerprint, so an old pass can never stand in for a new test.
Check it yourself
The evidence bundle holds this run's steps, checks and tools, the CLI record each step was judged against where it read one, and the published transcript. Its SHA-256 is listed in an index that is witnessed on a WARD chain and signed. Internal host names are replaced by their role.
Save the bundle, the verifier, index.json and index.receipt.json in one folder, then run node verify-journey-bundle.mjs ev_mw2tmmpc_2a98f4.json (Node 18 or later, offline, no account). It fails if a single byte has changed. A pass proves the bundle is the one we published, unchanged; the verdict inside it is our proof engine's, not an independent audit.