Lewati ke isi

T0-E1.1 independent evidence/validator audit

Manifest field audit

Field Manifest Independent observation Result
FE/BE commits present FE served identity matches; BE endpoint reports env SHA Partially verified
FE/BE tree hashes present Recomputed source hashes match current trees Verified for repository trees
FE build present Recomputed dist checksum and version match Verified
DB name/UUID present DB name/UUID match the live database Verified
Fixture version/checksum gt-core-e11 / null Live parameter is gt-core-b1.1; no checksum FAIL
Installed modules [] Live DB has 14 installed Scola modules FAIL
Entitlements {} Manifest does not contain a live entitlement payload/checksum source FAIL
Route denominator present Recomputed route checksum matches Verified
Required/executed cases 28/28 Set equality passes for credential result Syntactic only
Served BE source env commit + three fingerprints Commit is environment-derived; fingerprints cover only three files INSUFFICIENT

Validator weaknesses reproduced

  1. It checks module/entitlement checksum fields for presence but never recomputes their payloads from the live DB.
  2. It accepts golden_fixture_checksum=null because no non-empty fixture checksum is required.
  3. It accepts a fixture version string without comparing it to the live fixture state.
  4. It accepts backend identity sourced solely from SCOLA_BACKEND_GIT_COMMIT_SHA.
  5. It validates required-case set equality but does not independently validate each case's meaningful assertion or bind security-matrix results to the same denominator.
  6. The generated manifest can produce empty modules/entitlements and null fixture checksum when environment variables are absent; this is generation success, not fail-closed evidence generation.

These are sufficient for a false-green certification even though the validator's synthetic mutation tests pass.