Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Provides final code cleanup after task review approval. Removes debug logs, temporary comments, dead code, optimizes imports, and improves readability. Use when asked to clean up code, polish, finalize, tidy up, remove technical debt, or prepare code for completion after review. Not for refactoring logic or fixing bugs—focused solely on cosmetic and hygiene cleanup.
.claude/skills/giuseppe-trisciuoglio-specs-code-cleanup/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 606% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 48% | 0% |
Performs post-review cosmetic cleanup to make code production-ready. This workflow is now integrated as Phase T-7 of /developer-kit-specs:specs.task-implementation. It can also be invoked manually using --action=cleanup.
Input: docs/specs/[id]/tasks/TASK-XXX.md (reviewed status) Output: Cleaned code, task marked completed
| Argument | Required | Description | |----------|----------|-------------| | --lang | No | java, spring, typescript, nestjs, react, python, general | | --task | Yes | Path to task file | | --action| No | Set to cleanup for manual invocation |
./mvnw spotless:apply, npm run lint:fix, black, etc.See references/language-patterns.md for language-specific formatter commands, import ordering, and grep patterns.
$ARGUMENTS for parameters:--lang (optional): Target language/framework--task (required): Task ID or file path--spec (optional): Spec folder path (used with task ID)Support two formats:
--task=docs/specs/001-feature/tasks/TASK-001.md--spec=docs/specs/001-feature --task=TASK-001If Format 2 is used, construct the task file path as: {spec}/tasks/{task}.md
reviewed or implemented (not completed)TASK-XXX--review.md exists and is approved/developer-kit-specs:specs.task-review firstprovides filesTASK-XXX--review.md for files created/modifiedprovides field for file pathsSearch files for temporary/debug artifacts with Grep:
console.log, System.out.println, print(, // DEBUG:, // temp, // hackTODO/FIXME comments (keep unresolved ones)Review context for each finding. Remove confirmed debt and document what was removed.
## Cleanup Summary section to the task filecompleted and set completed_date + cleanup_date## Cleanup Summary to task file with:bash/developer-kit-specs:specs.task-implementation --lang=spring --task="docs/specs/001-user-auth/tasks/TASK-001.md" --action=cleanup
Actions:
reviewedUserController.java, UserService.java, UserRepository.javaSystem.out.println and 2 resolved TODOs./mvnw spotless:apply./mvnw test -qcompletedbash/developer-kit-specs:specs.task-implementation --lang=typescript --task="docs/specs/002-dashboard/tasks/TASK-003.md" --action=cleanup
Actions:
reviewedDashboard.tsx, useDashboard.ts, Dashboard.test.tsxconsole.log statementsnpm run lint:fix and npm run formatnpm testcompleted| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 7,260 | 5,676 | -22% | 1 | 1 | 0% | 257 | 1,705 | +563% | 0 | 0 | — |
case-01 | fail→fail | 5,324 | 3,339 | -37% | 1 | 1 | 0% | 274 | 1,765 | +544% | 0 | 0 | — |
case-18 | pass→fail | 12,284 | 2,864 | -77% | 1 | 1 | 0% | 1,606 | 1,802 | +12% | 0 | 0 | — |
case-03 | fail→fail | 5,815 | 6,067 | +4% | 1 | 1 | 0% | 237 | 1,687 | +612% | 0 | 0 | — |
case-04 | fail→pass | 13,886 | 6,398 | -54% | 1 | 1 | 0% | 2,494 | 2,383 | -4% | 0 | 0 | — |
case-05 | fail→fail | 9,662 | 5,715 | -41% | 1 | 1 | 0% | 1,726 | 1,824 | +6% | 0 | 0 | — |
case-06 | fail→pass | 6,716 | 4,280 | -36% | 1 | 1 | 0% | 294 | 2,076 | +606% | 0 | 0 | — |
case-07 | pass→pass | 8,189 | 2,992 | -63% | 1 | 1 | 0% | 1,487 | 1,909 | +28% | 0 | 0 | — |
case-08 | fail→pass | 8,709 | 3,416 | -61% | 1 | 1 | 0% | 1,400 | 1,922 | +37% | 0 | 0 | — |
case-09 | fail→pass | 9,949 | 4,300 | -57% | 1 | 1 | 0% | 1,517 | 2,080 | +37% | 0 | 0 | — |
case-10 | pass→pass | 6,194 | 2,574 | -58% | 1 | 1 | 0% | 1,062 | 1,787 | +68% | 0 | 0 | — |
case-11 | pass→pass | 5,029 | 5,036 | +0% | 1 | 1 | 0% | 740 | 2,233 | +202% | 0 | 0 | — |
case-12 | fail→pass | 8,433 | 2,725 | -68% | 1 | 1 | 0% | 1,251 | 1,850 | +48% | 0 | 0 | — |
case-13 | fail→pass | 9,491 | 3,855 | -59% | 1 | 1 | 0% | 1,505 | 1,951 | +30% | 0 | 0 | — |
case-14 | pass→pass | 9,764 | 3,294 | -66% | 1 | 1 | 0% | 1,471 | 1,881 | +28% | 0 | 0 | — |
case-15 | pass→pass | 9,918 | 6,610 | -33% | 1 | 1 | 0% | 1,702 | 2,478 | +46% | 0 | 0 | — |
case-16 | fail→pass | 8,152 | 4,474 | -45% | 1 | 1 | 0% | 1,287 | 2,129 | +65% | 0 | 0 | — |
case-17 | fail→pass | 22,508 | 2,309 | -90% | 1 | 1 | 0% | 2,367 | 1,752 | -26% | 0 | 0 | — |
case-19 | fail→pass | 6,768 | 1,661 | -75% | 1 | 1 | 0% | 976 | 1,604 | +64% | 0 | 0 | — |
case-20 | fail→pass | 10,723 | 2,595 | -76% | 1 | 1 | 0% | 1,566 | 1,774 | +13% | 0 | 0 | — |
case-21 | fail→pass | 11,656 | 4,843 | -58% | 1 | 1 | 0% | 1,911 | 2,108 | +10% | 0 | 0 | — |
case-22 | fail→pass | 8,907 | 1,851 | -79% | 1 | 1 | 0% | 1,382 | 1,606 | +16% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +50 percentage points is the difference between those two pass rates over the 18 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.