Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Full-sweep mode: runs a unified analysis across all quality dimensions — code decay, architecture, tech debt, and test quality — then applies fixes directly to the codebase. Safe changes are auto-applied; risky changes are confirmed before execution. Drawing on twelve classic...
.claude/skills/sickn33-brooks-sweep/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 454% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -40% | 0% |
Use this skill when you need full-sweep mode: runs a unified analysis across all quality dimensions — code decay, architecture, tech debt, and test quality — then applies fixes directly to the codebase. Safe changes are auto-applied; risky changes are confirmed before execution. Drawing on twelve classic...
../_shared/common.md for the Iron Law, Project Config, Report Template, and Health Score rules../_shared/source-coverage.md for book-level coverage, exceptions, and tradeoffs../_shared/decay-risks.md for production risk symptom definitions../_shared/test-decay-risks.md for test risk symptom definitionssweep-guide.md in this directory for the unified scan and fix processIf the user has not specified a project or directory: apply Auto Scope Detection from ../_shared/common.md to determine the review scope before proceeding.
unresolvable / non_critical_rounds / fix_log state (Step 1 of the guide)unresolvable set, cap non-critical rounds at 3 (Step 6 of the guide)Mode line in report: Full Sweep
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 1,267 | 6,155 | +386% | 1 | 1 | 0% | 203 | 1,125 | +454% | 0 | 0 | — |
case-02 | fail→fail | 7,930 | 19,192 | +142% | 1 | 1 | 0% | 1,278 | 794 | -38% | 0 | 0 | — |
case-03 | fail→fail | 10,288 | 3,179 | -69% | 1 | 1 | 0% | 1,844 | 930 | -50% | 0 | 0 | — |
case-04 | pass→fail | 6,669 | 5,563 | -17% | 1 | 1 | 0% | 1,143 | 1,539 | +35% | 0 | 0 | — |
case-05 | fail→pass | 9,078 | 2,418 | -73% | 1 | 1 | 0% | 1,606 | 920 | -43% | 0 | 0 | — |
case-06 | fail→pass | 17,917 | 2,892 | -84% | 1 | 1 | 0% | 1,433 | 1,072 | -25% | 0 | 0 | — |
case-07 | fail→pass | 7,755 | 3,081 | -60% | 1 | 1 | 0% | 1,277 | 1,051 | -18% | 0 | 0 | — |
case-08 | pass→pass | 7,353 | 3,536 | -52% | 1 | 1 | 0% | 1,315 | 1,103 | -16% | 0 | 0 | — |
case-09 | fail→pass | 11,953 | 3,708 | -69% | 1 | 1 | 0% | 1,900 | 1,141 | -40% | 0 | 0 | — |
case-10 | pass→pass | 8,826 | 3,113 | -65% | 1 | 1 | 0% | 1,540 | 1,079 | -30% | 0 | 0 | — |
case-11 | fail→pass | 21,625 | 1,589 | -93% | 1 | 1 | 0% | 1,580 | 739 | -53% | 0 | 0 | — |
case-12 | fail→pass | 9,334 | 6,438 | -31% | 1 | 1 | 0% | 1,540 | 1,533 | -0% | 0 | 0 | — |
case-13 | fail→pass | 12,325 | 7,268 | -41% | 1 | 1 | 0% | 2,064 | 1,763 | -15% | 0 | 0 | — |
case-14 | fail→pass | 6,544 | 2,237 | -66% | 1 | 1 | 0% | 1,234 | 870 | -29% | 0 | 0 | — |
case-15 | fail→pass | 8,353 | 1,789 | -79% | 1 | 1 | 0% | 1,459 | 845 | -42% | 0 | 0 | — |
case-16 | fail→fail | 11,130 | 1,538 | -86% | 1 | 1 | 0% | 1,902 | 754 | -60% | 0 | 0 | — |
case-17 | fail→pass | 6,064 | 3,549 | -41% | 1 | 1 | 0% | 1,004 | 1,054 | +5% | 0 | 0 | — |
case-18 | fail→fail | 6,741 | 4,425 | -34% | 1 | 1 | 0% | 1,153 | 1,255 | +9% | 0 | 0 | — |
case-19 | pass→pass | 3,110 | 2,725 | -12% | 1 | 1 | 0% | 593 | 1,035 | +75% | 0 | 0 | — |
case-20 | pass→pass | 9,005 | 7,077 | -21% | 1 | 1 | 0% | 1,838 | 1,970 | +7% | 0 | 0 | — |
case-21 | pass→pass | 11,894 | 9,075 | -24% | 1 | 1 | 0% | 2,384 | 2,338 | -2% | 0 | 0 | — |
case-22 | fail→fail | 8,945 | 1,754 | -80% | 1 | 1 | 0% | 1,474 | 825 | -44% | 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 20 counted toward the lift figure. The other 2 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 +45 percentage points is the difference between those two pass rates over the 20 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.