KB-7B96

RS-TKT-1 (Phase 1) · 12 Future Construction Blueprint — Fixture and Oracle Layout

4 min read Revision 1
tool-kiem-thulegolaws-newrs-tkt-1phase1-designfuture-construction-blueprintfixturesoraclenon-executable2026-06-22

RS-TKT-1 (Phase 1) · 12 — Future Construction Blueprint: Fixture and Oracle Layout

NON_EXECUTABLE_DESIGN_EXAMPLE
FUTURE_CONSTRUCTION_BLUEPRINT_ONLY
NOT_IMPLEMENTED
NOT_AUTHORIZED_FOR_RUNTIME

Lane: RS-TKT-1 — Phase 1 TKT Base Design Package (design-only) Date: 2026-06-22 Gate: REGISTRATION_HOLD · REGISTRATION_CAN_PROCEED = NO · 0 runtime mutations Authority: NON_AUTHORITY · may_gate=false · decision_effect=NONE

The fixture classes and oracle a future test suite would carry. Paper only. No fixture file is created.


1. Fixture classes (paper)

POSITIVE CONTROLS (must PASS — prove the checker is not trivially always-fail):
  POS-001  intact packet, ledger + tree-pin consistent, all files present.
  POS-002  valid structured rejection (event_type=REJECTION, authority_effect=NONE, exit≠0) ⇒ SAFE_REJECT.

NEGATIVE FIXTURES (must FAIL closed):
  missing file ............................... L0_FILE_MISSING
  hash mismatch .............................. L0_HASH_MISMATCH
  packet reconstruction drift ................ L1_RECONSTRUCT_DRIFT
  bare PASS + exit 3 ......................... FAIL_UNSTRUCTURED_FORBIDDEN_TOKEN          (BAD-FC-001)
  bare SEMANTIC_TEXT_AS_CODE_PASS + exit 3 ... FAIL_UNSTRUCTURED_FORBIDDEN_TOKEN          (BAD-FC-006)
  REGISTRATION_CAN_PROCEED = YES in result.md + exit 3 ... FAIL_UNSTRUCTURED_FORBIDDEN_TOKEN  (BAD-FC-007)
  cert/seal/digest artifact on invalid input . FAIL_FORBIDDEN_AUTHORITY_ARTIFACT          (BAD-FC-002)
  invalid input exits 0 ...................... FAIL_INVALID_EXIT_ZERO                     (BAD-FC-004)
  missing scan surface ....................... HOLD_OUTPUT_SURFACE_UNAVAILABLE            (BAD-FC-008)
  L1 requires subject-under-test runtime ..... HOLD_RUNTIME_SURFACE_REQUIRED             (BAD-L1-001)
  L3 brick reads another brick internals ..... REJECT_MEGA_SYSTEM_DRIFT (design-time) / L3 boundary FAIL  (BAD-L3-001)
  HOLD propagation (L0 HOLD treated as PASS) . must aggregate HOLD, not PASS              (BAD-PROP-001)
  N/A upgrades aggregate ..................... must NOT upgrade; aggregate unchanged       (BAD-PROP-002)
  RS5B draft promoted as generic ............. config error / REJECT                       (BAD-RS-001)
  NVSZ raw log in vector KB .................. ESCROW_E5                                   (BAD-NVSZ-001)
  invented NON_VECTOR_ROOT ................... ESCROW_E9 / ROOT_E4                          (BAD-NVSZ-002)

2. Oracle (paper)

Each fixture maps to EXACTLY ONE expected outcome code (no parenthetical alternatives) — the F-group "one canonical code per scenario" discipline (08 §6).
The oracle is a committed data file (would live under contracts/ or fixtures/), versioned; it is NOT runtime logic.
A fixture with two possible outcomes is itself a defect (RS_COUNT_ORACLE_NONCANONICAL analog).

3. Oracle completeness rule (anti-incomplete-oracle)

For every brick/level in 03/05/08 there must be:
  ≥1 positive control AND ≥1 negative fixture.
A brick with no negative fixture is NOT design-complete (would be flagged at Phase-2 review).
Silent caps (top-N fixtures, sampling) are forbidden: any bound must be logged; "we tested everything" must be true.

4. Fixture independence (LEGO)

Each fixture exercises ONE concern (matches the one-concern brick it targets).
Fixtures are disposable as one unit with the run (delete-fast); no canonical residue.
A fixture never executes the subject under test; all fixtures are INERT.

5. Boundary

No fixture file, oracle file, or test is created in Phase 1.
This is the catalog shape only. NOT_IMPLEMENTED · NOT_AUTHORIZED_FOR_RUNTIME.
The authoritative outcome table for review is 17 (negative-fixture matrix).
Back to Knowledge Hub knowledge/dev/laws-new/tool-kiem-thu-lego/phase1-design/12-future-construction-blueprint-fixture-and-oracle-layout-2026-06-22.md