Harden review instructions against self-supplied evidence

.2119/verdicts/REQ-003.2.2--bea345f6b733.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.2.2--bea345f6b733",
+  "requirementId": "REQ-003.2.2",
+  "hash": "bea345f6b733",
+  "verdict": "pass",
+  "summary": "pass and fail serialize verdict records as plain JSON in .2119/verdicts/, while init and verdict writes repair ignore rules so those records remain trackable.",
+  "timestamp": "2026-08-02T22:29:35.122Z"
+}

.2119/verdicts/REQ-003.8.1--683e7215f03e.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.1--683e7215f03e",
+  "requirementId": "REQ-003.8.1",
+  "hash": "683e7215f03e",
+  "verdict": "fail",
+  "summary": "eval/calibration's 13 cases omit the repository's documented task_lib.sh-to-'shell libraries' verdict-scope inflation escape, so the corpus no longer contains every known review escape.",
+  "timestamp": "2026-08-02T22:32:53.245Z"
+}

.2119/verdicts/REQ-003.8.1--87a0d11bd0c2.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.1--87a0d11bd0c2",
+  "requirementId": "REQ-003.8.1",
+  "hash": "87a0d11bd0c2",
+  "verdict": "pass",
+  "summary": "The corpus fixtures retain the required fields and documented escapes, while corrected case 013 now accurately condenses missing/invalid verdict rejection and real-CLI acceptance from its cited production test.",
+  "timestamp": "2026-08-02T23:02:34.369Z"
+}

.2119/verdicts/REQ-003.8.1--b858103eb35c.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.1--b858103eb35c",
+  "requirementId": "REQ-003.8.1",
+  "hash": "b858103eb35c",
+  "verdict": "fail",
+  "summary": "The listed fixtures have the required fields, but 'every known review escape' has no defined external inventory or scope, so corpus completeness is ambiguous and cannot be verified from the corpus's self-claim.",
+  "timestamp": "2026-08-02T22:31:09.428Z"
+}

.2119/verdicts/REQ-003.8.1--c355ec3bc765.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.1--c355ec3bc765",
+  "requirementId": "REQ-003.8.1",
+  "hash": "c355ec3bc765",
+  "verdict": "pass",
+  "summary": "The calibration corpus contains the required fixture fields and now includes 014 for the documented PR #112 singular-evidence-to-plural-category scope-inflation escape.",
+  "timestamp": "2026-08-02T22:34:46.490Z"
+}

.2119/verdicts/REQ-003.8.2--5262e7c71e7b.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.2--5262e7c71e7b",
+  "requirementId": "REQ-003.8.2",
+  "hash": "5262e7c71e7b",
+  "verdict": "fail",
+  "summary": "The current template cannot preserve calibration case 013's expected PASS: its conjunct rule requires rejected counterexamples for empty note, mismatched owner, and unparseable stamp, which that case's evidence does not provide.",
+  "timestamp": "2026-08-02T22:33:18.373Z"
+}

.2119/verdicts/REQ-003.8.2--5ec997925e8d.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.2--5ec997925e8d",
+  "requirementId": "REQ-003.8.2",
+  "hash": "5ec997925e8d",
+  "verdict": "fail",
+  "summary": "Calibration case 013-pass-control-negative-positive expects PASS, but the template requires a rejected counterexample for every conjunct while its evidence never independently rejects an empty note, mismatched owner, or unparseable stamp; a reviewer following the template must FAIL that case.",
+  "timestamp": "2026-08-02T22:31:16.326Z"
+}

.2119/verdicts/REQ-003.8.2--d52ce5a3aa84.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.2--d52ce5a3aa84",
+  "requirementId": "REQ-003.8.2",
+  "hash": "d52ce5a3aa84",
+  "verdict": "pass",
+  "summary": "The current standard and audit instructions retain the checks needed by calibration cases 001-014: tautology/over-mocking flags, provenance, conjunct and boundary counterexamples, requirement-quality judgment, and evidence-bounded summary scope; case 013 now rejects each named malformed-record conjunct and keeps a genuine-writer pass control.",
+  "timestamp": "2026-08-02T22:34:56.716Z"
+}

.2119/verdicts/REQ-003.8.2--deb61e2892be.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.2--deb61e2892be",
+  "requirementId": "REQ-003.8.2",
+  "hash": "deb61e2892be",
+  "verdict": "fail",
+  "summary": "The test-quality template forbids PASS without production file:line provenance, but calibration cases 012-pass-control-perturbation and 013-pass-control-negative-positive contain only condensed snippets and expect PASS, so a reviewer following the template cannot preserve those two corpus verdicts.",
+  "timestamp": "2026-08-02T22:29:50.495Z"
+}

.2119/verdicts/REQ-003.8.2--ec2046ab17a0.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-003.8.2--ec2046ab17a0",
+  "requirementId": "REQ-003.8.2",
+  "hash": "ec2046ab17a0",
+  "verdict": "pass",
+  "summary": "The current standard, direct-judgment, and audit instructions preserve the expected verdicts for calibration cases 001-014; corrected case 013 now independently covers its missing verdict, invalid verdict, and real-CLI acceptance clauses.",
+  "timestamp": "2026-08-02T23:02:58.593Z"
+}

.2119/verdicts/REQ-004.1.3--f46aea2e9d6b.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-004.1.3--f46aea2e9d6b",
+  "requirementId": "REQ-004.1.3",
+  "hash": "f46aea2e9d6b",
+  "verdict": "fail",
+  "summary": "The tests handcraft hook payloads instead of obtaining a production platform payload, omit Codex Delete File apply_patch paths and custom/test-glob positive boundaries, and direct helper calls do not prove path determination occurs inside the CLI.",
+  "timestamp": "2026-08-02T22:31:15.380Z"
+}

.2119/verdicts/REQ-004.1.7--0cff14d8a0f3.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-004.1.7--0cff14d8a0f3",
+  "requirementId": "REQ-004.1.7",
+  "hash": "0cff14d8a0f3",
+  "verdict": "fail",
+  "summary": "The annotated test calls handleHook directly and never exercises generated platform hook wiring; deleting the SessionStart entries in src/adapters.ts would prevent session-start injection through the platform mechanism while every REQ-004.1.7 assertion stays green.",
+  "timestamp": "2026-08-02T22:33:04.234Z"
+}

.2119/verdicts/REQ-004.1.7--69cfdfd5ebdd.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-004.1.7--69cfdfd5ebdd",
+  "requirementId": "REQ-004.1.7",
+  "hash": "69cfdfd5ebdd",
+  "verdict": "fail",
+  "summary": "The test does not execute the installed SessionStart command: a broken command can retain the asserted hook substring while the separately invoked dist/cli.js still returns context, leaving the platform integration broken.",
+  "timestamp": "2026-08-02T22:36:05.854Z"
+}

.2119/verdicts/REQ-004.1.7--9247f44ae9d9.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-004.1.7--9247f44ae9d9",
+  "requirementId": "REQ-004.1.7",
+  "hash": "9247f44ae9d9",
+  "verdict": "fail",
+  "summary": "The test separately inspects installed SessionStart command text and calls handleHook, but never executes that command through the CLI; breaking src/cli.ts hook dispatch would prevent all three installed commands from injecting context while every annotated assertion stays green.",
+  "timestamp": "2026-08-02T22:34:39.173Z"
+}

.2119/verdicts/REQ-004.1.7--a4801c86dd0e.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-004.1.7--a4801c86dd0e",
+  "requirementId": "REQ-004.1.7",
+  "hash": "a4801c86dd0e",
+  "verdict": "fail",
+  "summary": "The test covers only Claude and derives the expected body from production SESSION_CONTEXT; Codex/Gemini omission or a non-descriptive/overlong context can remain green, while 'short' has no testable bound.",
+  "timestamp": "2026-08-02T22:29:38.198Z"
+}

.2119/verdicts/REQ-004.1.7--a6697c983bee.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "REQ-004.1.7--a6697c983bee",
+  "requirementId": "REQ-004.1.7",
+  "hash": "a6697c983bee",
+  "verdict": "pass",
+  "summary": "For each installed Claude, Codex, and Gemini SessionStart command, the test executes that production command and requires successful JSON with SessionStart additionalContext naming spec-driven testing, MUST-level obligations, and the check command in under 1000 characters.",
+  "timestamp": "2026-08-02T22:37:23.963Z"
+}

.2119/verdicts/self-supplied-evidence.1.1--3a42292f6705.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.1--3a42292f6705",
+  "requirementId": "self-supplied-evidence.1.1",
+  "hash": "3a42292f6705",
+  "verdict": "pass",
+  "summary": "Every production-computed test-quality task is compared against the complete canonical task body, so removing, weakening, or contradicting the concrete-production-failure mandate fails the test.",
+  "timestamp": "2026-08-02T23:10:57.874Z"
+}

.2119/verdicts/self-supplied-evidence.1.1--6419602f2307.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.1--6419602f2307",
+  "requirementId": "self-supplied-evidence.1.1",
+  "hash": "6419602f2307",
+  "verdict": "fail",
+  "summary": "The exact mandate is asserted, but the anti-waiver check misses contradictory guidance such as 'A generic category-level failure is sufficient,' which permits a non-concrete answer while the test stays green.",
+  "timestamp": "2026-08-02T23:07:14.248Z"
+}

.2119/verdicts/self-supplied-evidence.1.1--7c4774226fe5.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.1--7c4774226fe5",
+  "requirementId": "self-supplied-evidence.1.1",
+  "hash": "7c4774226fe5",
+  "verdict": "pass",
+  "summary": "Every production-identified test-quality instruction is required to ask for a concrete production failure, while tests reject omission, optionality, and contradictory carve-outs.",
+  "timestamp": "2026-08-02T05:32:14.238Z"
+}

.2119/verdicts/self-supplied-evidence.1.1--8e33566ca8c8.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.1--8e33566ca8c8",
+  "requirementId": "self-supplied-evidence.1.1",
+  "hash": "8e33566ca8c8",
+  "verdict": "fail",
+  "summary": "The provenance slice is exact, but contradictory guidance appended after the counterexample block can permit a generic failure category while every 1.1 assertion remains green.",
+  "timestamp": "2026-08-02T23:08:33.884Z"
+}

.2119/verdicts/self-supplied-evidence.1.1--ef8087c8df76.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.1--ef8087c8df76",
+  "requirementId": "self-supplied-evidence.1.1",
+  "hash": "ef8087c8df76",
+  "verdict": "pass",
+  "summary": "Every production-computed test-quality instruction retains the complete canonical task body, so removing, weakening, or contradicting the concrete-production-failure mandate fails the test.",
+  "timestamp": "2026-08-02T23:14:00.257Z"
+}

.2119/verdicts/self-supplied-evidence.1.2--419abd7dee19.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.2--419abd7dee19",
+  "requirementId": "self-supplied-evidence.1.2",
+  "hash": "419abd7dee19",
+  "verdict": "pass",
+  "summary": "Every production-discovered test-quality instruction is compared against the complete expected task body with only dynamic IDs normalized, so removing, weakening, or contradicting the file:line independence demand fails.",
+  "timestamp": "2026-08-02T23:10:46.134Z"
+}

.2119/verdicts/self-supplied-evidence.1.2--4f90050b0b09.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.2--4f90050b0b09",
+  "requirementId": "self-supplied-evidence.1.2",
+  "hash": "4f90050b0b09",
+  "verdict": "pass",
+  "summary": "Every production-discovered test-quality instruction is still matched against the complete expected task with only dynamic IDs normalized, so any omission, weakening, or contradiction of the file:line production-reachability demand fails.",
+  "timestamp": "2026-08-02T23:14:09.042Z"
+}

.2119/verdicts/self-supplied-evidence.1.2--6331058954ea.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.2--6331058954ea",
+  "requirementId": "self-supplied-evidence.1.2",
+  "hash": "6331058954ea",
+  "verdict": "pass",
+  "summary": "All generated test-quality tasks are exhaustively selected from production review targets and require file:line production reachability for the named failure, with exact independence from test, fixture, and prompt-supplied triggers or decisive observations and explicit rejection of opt-outs.",
+  "timestamp": "2026-08-02T05:32:24.495Z"
+}

.2119/verdicts/self-supplied-evidence.1.2--78804c69a5b9.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.2--78804c69a5b9",
+  "requirementId": "self-supplied-evidence.1.2",
+  "hash": "78804c69a5b9",
+  "verdict": "fail",
+  "summary": "The test stays green if the provenance section retains the exact mandate but adds 'Evidence from test fixtures satisfies this demand.'; that unrecognized contradiction permits fixture-supplied evidence despite 1.2's independence clause.",
+  "timestamp": "2026-08-02T23:07:12.165Z"
+}

.2119/verdicts/self-supplied-evidence.1.2--82dcac5dc17c.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.1.2--82dcac5dc17c",
+  "requirementId": "self-supplied-evidence.1.2",
+  "hash": "82dcac5dc17c",
+  "verdict": "fail",
+  "summary": "The exact comparison ends at '**Counterexample obligation:**'; appending later in the same instruction 'Evidence from test fixtures satisfies the production-provenance demand.' leaves every assertion green while negating 1.2's independence requirement.",
+  "timestamp": "2026-08-02T23:08:31.147Z"
+}

.2119/verdicts/self-supplied-evidence.2.1--1fc1ddf5c224.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.1--1fc1ddf5c224",
+  "requirementId": "self-supplied-evidence.2.1",
+  "hash": "1fc1ddf5c224",
+  "verdict": "pass",
+  "summary": "The real review-dispatch CLI generates every computed test-quality instruction, and the test rejects omission or alteration of consumption, emitted value, separate invocation, production component, or production data-source terms as well as weakening exceptions.",
+  "timestamp": "2026-08-02T23:07:04.908Z"
+}

.2119/verdicts/self-supplied-evidence.2.1--2aa103ac58bd.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.1--2aa103ac58bd",
+  "requirementId": "self-supplied-evidence.2.1",
+  "hash": "2aa103ac58bd",
+  "verdict": "pass",
+  "summary": "All generated test-quality tasks are enumerated from production review targets and must contain the exact narrow boundary definition; omission or alteration of consumption, emitted value, separate invocation, or production component/data-source terms fails, while anti-waiver assertions reject contradictory weakening.",
+  "timestamp": "2026-08-02T05:32:29.175Z"
+}

.2119/verdicts/self-supplied-evidence.2.1--611ec8687d7d.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.1--611ec8687d7d",
+  "requirementId": "self-supplied-evidence.2.1",
+  "hash": "611ec8687d7d",
+  "verdict": "pass",
+  "summary": "Every computed production-generated test-quality task is compared with a complete oracle after only dynamic ID normalization, so changing consumption, emitted value, separate invocation, production-component, or production-data-source scope fails.",
+  "timestamp": "2026-08-02T23:14:10.359Z"
+}

.2119/verdicts/self-supplied-evidence.2.1--7da77538aed0.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.1--7da77538aed0",
+  "requirementId": "self-supplied-evidence.2.1",
+  "hash": "7da77538aed0",
+  "verdict": "pass",
+  "summary": "Every production-dispatched test-quality instruction must match an independent full provenance oracle, so altering any producer/consumer boundary conjunct—consumption, emitted value, separate invocation, production component, or production data source—fails.",
+  "timestamp": "2026-08-02T23:08:26.294Z"
+}

.2119/verdicts/self-supplied-evidence.2.1--a000c7a0aa4b.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.1--a000c7a0aa4b",
+  "requirementId": "self-supplied-evidence.2.1",
+  "hash": "a000c7a0aa4b",
+  "verdict": "pass",
+  "summary": "Every generated test-quality task is matched against a complete independent task oracle after normalizing only requirement/review IDs, so any change to the producer/consumer definition's consumption, emitted value, separate invocation, production component, or data-source scope fails.",
+  "timestamp": "2026-08-02T23:10:44.573Z"
+}

.2119/verdicts/self-supplied-evidence.2.2--156362cfdb04.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.2--156362cfdb04",
+  "requirementId": "self-supplied-evidence.2.2",
+  "hash": "156362cfdb04",
+  "verdict": "fail",
+  "summary": "The assertion checks only the citation clause substring: changing its trigger to 'Unless that boundary exists' leaves the test and anti-waiver regexes green while no longer requiring evidence whenever a producer/consumer boundary exists.",
+  "timestamp": "2026-08-02T23:07:37.188Z"
+}

.2119/verdicts/self-supplied-evidence.2.2--56dbe92c6003.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.2--56dbe92c6003",
+  "requirementId": "self-supplied-evidence.2.2",
+  "hash": "56dbe92c6003",
+  "verdict": "pass",
+  "summary": "Every production-computed test-quality task pins the producer/consumer condition and requires file:line evidence that the covering test obtains its input from that production producer.",
+  "timestamp": "2026-08-02T23:14:23.989Z"
+}

.2119/verdicts/self-supplied-evidence.2.2--8fa7baeaefb2.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.2--8fa7baeaefb2",
+  "requirementId": "self-supplied-evidence.2.2",
+  "hash": "8fa7baeaefb2",
+  "verdict": "pass",
+  "summary": "The test inspects every generated test-quality task, requires mandatory file:line evidence that its input comes from the production producer, and rejects waiver or self-supplied-evidence language.",
+  "timestamp": "2026-08-02T05:33:28.165Z"
+}

.2119/verdicts/self-supplied-evidence.2.2--a3e153ca68e4.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.2--a3e153ca68e4",
+  "requirementId": "self-supplied-evidence.2.2",
+  "hash": "a3e153ca68e4",
+  "verdict": "pass",
+  "summary": "Each generated test-quality task must match the complete normalized task oracle, which pins the existing-boundary trigger, file:line citation, covering test input, and production producer together and rejects reversal, omission, or weakening.",
+  "timestamp": "2026-08-02T23:11:02.928Z"
+}

.2119/verdicts/self-supplied-evidence.2.2--df4ea1367363.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.2--df4ea1367363",
+  "requirementId": "self-supplied-evidence.2.2",
+  "hash": "df4ea1367363",
+  "verdict": "pass",
+  "summary": "The independent full-block oracle now pins 'If that boundary exists' together with file:line citation, test input, and that producer for every production-dispatched test-quality instruction, rejecting reversed or weakened triggers.",
+  "timestamp": "2026-08-02T23:08:52.297Z"
+}

.2119/verdicts/self-supplied-evidence.2.3--3bfc1f157e4c.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.3--3bfc1f157e4c",
+  "requirementId": "self-supplied-evidence.2.3",
+  "hash": "3bfc1f157e4c",
+  "verdict": "pass",
+  "summary": "The generated-task test covers every production-derived test-quality target and rejects omission or weakening of the required file:line trace that the exercised input preserves the production producer's value shape.",
+  "timestamp": "2026-08-02T05:33:56.135Z"
+}

.2119/verdicts/self-supplied-evidence.2.3--a0e14d8db077.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.3--a0e14d8db077",
+  "requirementId": "self-supplied-evidence.2.3",
+  "hash": "a0e14d8db077",
+  "verdict": "pass",
+  "summary": "Every production-discovered test-quality task is full-body matched, pinning the conditional file:line demand that the covering test's exercised input preserve the production producer's value shape.",
+  "timestamp": "2026-08-02T23:14:29.592Z"
+}

.2119/verdicts/self-supplied-evidence.2.3--b768f1b9b787.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.3--b768f1b9b787",
+  "requirementId": "self-supplied-evidence.2.3",
+  "hash": "b768f1b9b787",
+  "verdict": "pass",
+  "summary": "Every production-discovered test-quality instruction is full-body matched, including the conditional file:line demand that the exercised value preserve the production producer's shape; omissions, weakenings, and contradictions fail.",
+  "timestamp": "2026-08-02T23:11:13.182Z"
+}

.2119/verdicts/self-supplied-evidence.2.4--2755a46ec65c.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.4--2755a46ec65c",
+  "requirementId": "self-supplied-evidence.2.4",
+  "hash": "2755a46ec65c",
+  "verdict": "pass",
+  "summary": "Fresh dispatch integration verifies every generated test-quality task explicitly requires file:line evidence distinguishing a newly produced observation from an equal pre-existing sentinel, while shared rejection checks forbid weakening language.",
+  "timestamp": "2026-08-02T05:34:49.897Z"
+}

.2119/verdicts/self-supplied-evidence.2.4--2e7bb01145d0.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.4--2e7bb01145d0",
+  "requirementId": "self-supplied-evidence.2.4",
+  "hash": "2e7bb01145d0",
+  "verdict": "pass",
+  "summary": "Every computed production-generated test-quality task matches the complete oracle, pinning the decisive-observation trigger, all four initial/default/placeholder/sentinel cases, file:line evidence, and newly-produced versus pre-existing distinction.",
+  "timestamp": "2026-08-02T23:14:34.728Z"
+}

.2119/verdicts/self-supplied-evidence.2.4--3add2f8a8874.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.2.4--3add2f8a8874",
+  "requirementId": "self-supplied-evidence.2.4",
+  "hash": "3add2f8a8874",
+  "verdict": "pass",
+  "summary": "The complete task snapshot pins the conditional trigger, all four pre-existing-value classes, file:line evidence, and the newly-produced-versus-pre-existing distinction for every production-computed test-quality target.",
+  "timestamp": "2026-08-02T23:11:20.691Z"
+}

.2119/verdicts/self-supplied-evidence.3.1--0c0cd9b96a10.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.3.1--0c0cd9b96a10",
+  "requirementId": "self-supplied-evidence.3.1",
+  "hash": "0c0cd9b96a10",
+  "verdict": "pass",
+  "summary": "Every production-computed test-quality task exactly defines the boundary as invocation of a binary or service outside the gate's own process, with the complete task snapshot rejecting weakened definitions.",
+  "timestamp": "2026-08-02T23:14:48.392Z"
+}

.2119/verdicts/self-supplied-evidence.3.1--8f4768f6eda9.json

@@ -0,0 +1,8 @@
+{
+  "reviewId": "self-supplied-evidence.3.1--8f4768f6eda9",
+  "requirementId": "self-supplied-evidence.3.1",
+  "hash": "8f4768f6eda9",
+  "verdict": "pass",
+  "summary": "Enumerates every production-derived test-quality task and rejects missing or altered boundary wording, including inside-process and non-binary/service near-counterexamples, via an exact required definition.",
+  "timestamp": "2026-08-02T05:36:17.365Z"
+}