Sutura Case Lab

JavaScript repair

Recorded live result Fixed

Passed the required checks

The repair reproduced the failure, fixed it, and held the behavior the checks require.

Historical recorded result · A real off-by-one bug fails one test. Sutura must repair the code, not the test. · subject 3b6152d9d72beebd7dcbaaa4045c94d640456adb · 2026-10-27T09:39:05.311Z

Review the diff and the behavior tested before merging.

This result is a recorded live benchmark evaluation of the released Sutura version, not a run started from this page.

Outcome matches the expected outcome.

Failed commit and CI evidence

Placebo case repair-off-by-one: An exact page boundary returns one extra page. The repair must change page-count.js and keep the assertion.
Failing command: vitest run
Error excerpt: [41m[1m FAIL [22m[49m case.test.js[2m > [22mkeeps an exact final page [31m[1mAssertionError[22m: expected 3 to be 2 // Object.is equality[39m [32m- Expected[39m [31m+ Received[39m [32m- 2[39m [31m+ 3[39m [36m [2m❯[22m case.test.js:[2m4:67[22m[39m [90m 2|[39m [35mimport[39m { pageCount } [35mfrom[39m [32m'./page-count.js'[39m[33m;[39m [90m 3|[39m [90m 4|[39m test('keeps an exact final page', () => expect(pageCount(20, 10)).toBe… [90m |[39m [31m^[39m [90m 5|[39m [31m[2m⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯[1/1]⎯[22m[39m [ELIFECYCLE] Test failed. See above for more details.
Repository local/fixture · run local-fixture · runtime node
Recorded workflow run

Nano diagnosis and confidence

Failure class

test-assertion

Confidence

49%

Signals

AssertionError: expected 3 to be 2 // Object.is equality
expected 3 to be 2
Object.is equality
mechanical:test-assertion
signature:\bexpected\b[^\n]*\b(?:to|but)\b[^\n]*\b(?:received|be|equal|match)\b
llm-command-mismatch
llm-excerpt-ungrounded

Diagnosis recovery

Initial diagnosis retained: test-assertion. Recovery: not-run — no-supported-recovery-signal.
Observed command: vitest run
Executed command: if [ -f pnpm-lock.yaml ]; then corepack pnpm exec sh -c 'vitest run'; elif [ -f yarn.lock ]; then corepack yarn exec sh -c 'vitest run'; else PATH="./node_modules/.bin:$PATH" sh -c 'vitest run'; fi
These bounded observations do not measure live repair quality.

Clean audit branch and Ultra verdict

Verdict: approved · policy default at local

Check	Result	Evidence
`deleted-test`	passed
`skipped-test`	passed
`pass-with-no-tests`	passed
`weakened-assertion`	passed
`loosened-type`	passed
`relaxed-config`	passed
`module-syntax`	passed
`llm-adjudication`	passed	The patch replaces the incorrect pageCount implementation (floor division plus one) with the correct ceiling division (with a zero-items guard). This directly addresses the assertion failure where pageCount(20, 10) returned 3 instead of the expected 2. The test suite passes after the change, confirming the fix without any reduction in test coverage or enforcement.
`second-opinion`	passed	gpt-6-astra: skipped: Skipped: second-opinion budget exhausted
`typesafe-audit`	passed	jev-1.14.0: approved: P(green-wash)=0.01 confidence=0.98; touches-only-tests=0.01 weakens-assertion=0.09 suppresses-error=0.03 unrelated-change=0.05
`policy-required-command`	passed	Passed 0 repository policy commands

The patch replaces the incorrect pageCount implementation (floor division plus one) with the correct ceiling division (with a zero-items guard). This directly addresses the assertion failure where pageCount(20, 10) returned 3 instead of the expected 2. The test suite passes after the change, confirming the fix without any reduction in test coverage or enforcement.

Final outcome

Fixed Recorded live result

Sutura never merges a generated repair. A fixed result still needs human review.

Check a patch this page did not produce

This run did not record an exact failing commit and policy commit, so both are shown as placeholders. Verification is always tied to exact commits. A patch cannot ask for a more permissive policy by carrying one, and no green log is accepted in place of execution.

sutura verify \
  --case-dir <checkout> \
  --source-sha <failing commit> \
  --policy-base-sha <trusted policy commit> \
  --candidate-diff <patch file> \
  --failing-command diagnosed \
  --format json

The same route runs as a GitHub Action and needs no repository write access. Uploading a patch stays a command-line and Action route; this page never runs one.

Execution and cost detail

Release: v0.4.1 · Action 3b6152d9d72beebd7dcbaaa4045c94d640456adb
Controller: 3b6152d9d72beebd7dcbaaa4045c94d640456adb
Request: recorded-javascript-repair · 2026-10-30T06:58:46.078Z

ConTree search tree and branch status

Triage: real · reproduced 4/4 · stop reason failure-boundary · method sprt-p20-p80-a05-b05-v1

Node	Parent	Depth	Terminal reason	Test exit	Policy	Changed files	Diff bytes
`search-001`	root	1	passed	0	valid	1	300
`search-002`	root	1	passed	0	valid	1	300
`search-003`	root	1	cancelled	1	valid	0	0

Super proposals and candidate patches

Candidate `repair-9b12d1fedabe` · selected

Replace the controller-selected source excerpt.

Test exit 0 · held its result · node search-001 · Adaptive search passed at depth 1

diff --git a/page-count.js b/page-count.js
index 28f73e0..22b843b 100644
--- a/page-count.js
+++ b/page-count.js
@@ -1 +1 @@
-export function pageCount(items, size) { return Math.floor(items / size) + 1; }
+export function pageCount(items, size) { return items === 0 ? 0 : Math.ceil(items / size); }

Candidate `repair-9b12d1fedabe` · selected

Replace the controller-selected source excerpt.

Test exit 0 · held its result · node search-002 · Adaptive search passed at depth 1

diff --git a/page-count.js b/page-count.js
index 28f73e0..22b843b 100644
--- a/page-count.js
+++ b/page-count.js
@@ -1 +1 @@
-export function pageCount(items, size) { return Math.floor(items / size) + 1; }
+export function pageCount(items, size) { return items === 0 ? 0 : Math.ceil(items / size); }

Rejected patches and rejection reasons

No candidate was rejected.

Token cost, provider cost, latency, and sandbox operations

Inference cost: USD 0.040622
Sandbox cost: USD 0.079354
Elapsed: 105.5 s
CPU time: 7.30 s
Peak memory: 144 MiB
Sandbox stages: 24 (13 with an operation id)

Role	Calls	Input tokens	Output tokens	Cost
nano	1	1846	471	USD 0.000224
super	3	3543	88	USD 0.001142
ultra	5	14428	535	USD 0.039256

Links

Recorded from docs/demo/placebo-v0.4.1-live-2026-10-27.json (result hash d8eb9f18c578…) at 2026-10-27T09:39:05.311Z, subject 3b6152d9d72beebd7dcbaaa4045c94d640456adb.

Result hash 73836cf0d6420ff7f5690cc5190ce578341164407f0dc77eb4d3d6a9c58dcd03

Passed the required checks

Failed commit and CI evidence

Nano diagnosis and confidence

Diagnosis recovery

Clean audit branch and Ultra verdict

Final outcome

Check a patch this page did not produce

ConTree search tree and branch status

Super proposals and candidate patches

Candidate repair-9b12d1fedabe · selected

Candidate repair-9b12d1fedabe · selected

Rejected patches and rejection reasons

Token cost, provider cost, latency, and sandbox operations

Links

Candidate `repair-9b12d1fedabe` · selected

Candidate `repair-9b12d1fedabe` · selected