Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use to run the pre-merge pull request review checklist so required validations are never skipped. Use when reviewing or approving a pull request before merge. Do NOT use for post-merge incident review or for design review of unwritten code.
.claude/skills/bonaniibm-pr-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -61% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | -53% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | -53% | 0% |
Run before approving any pull request for merge.
A PR adds a new endpoint but no tests. Block on the missing-test item and request coverage before approving.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,912 | 4,631 | -22% | 1 | 1 | 0% | 1,010 | 923 | -9% | 0 | 0 | — |
case-02 | fail→fail | 14,863 | 13,697 | -8% | 1 | 1 | 0% | 1,193 | 1,310 | +10% | 0 | 0 | — |
case-03 | fail→pass | 6,895 | 5,156 | -25% | 1 | 1 | 0% | 1,155 | 981 | -15% | 0 | 0 | — |
case-04 | pass→pass | 8,720 | 2,632 | -70% | 1 | 1 | 0% | 1,335 | 630 | -53% | 0 | 0 | — |
case-05 | pass→pass | 9,162 | 3,024 | -67% | 1 | 1 | 0% | 1,475 | 693 | -53% | 0 | 0 | — |
case-06 | pass→pass | 6,176 | 2,865 | -54% | 1 | 1 | 0% | 1,024 | 622 | -39% | 0 | 0 | — |
case-07 | pass→pass | 5,130 | 3,512 | -32% | 1 | 1 | 0% | 871 | 753 | -14% | 0 | 0 | — |
case-08 | pass→pass | 3,623 | 2,751 | -24% | 1 | 1 | 0% | 660 | 667 | +1% | 0 | 0 | — |
case-09 | pass→pass | 11,149 | 2,856 | -74% | 1 | 1 | 0% | 1,780 | 564 | -68% | 0 | 0 | — |
case-10 | pass→pass | 8,101 | 3,125 | -61% | 1 | 1 | 0% | 1,358 | 679 | -50% | 0 | 0 | — |
case-15 | pass→pass | 8,384 | 3,138 | -63% | 1 | 1 | 0% | 1,363 | 657 | -52% | 0 | 0 | — |
case-11 | pass→pass | 8,391 | 3,642 | -57% | 1 | 1 | 0% | 1,459 | 788 | -46% | 0 | 0 | — |
case-12 | fail→pass | 11,345 | 3,272 | -71% | 1 | 1 | 0% | 1,721 | 667 | -61% | 0 | 0 | — |
case-13 | fail→pass | 6,374 | 2,895 | -55% | 1 | 1 | 0% | 946 | 550 | -42% | 0 | 0 | — |
case-14 | pass→pass | 10,752 | 2,214 | -79% | 1 | 1 | 0% | 1,907 | 482 | -75% | 0 | 0 | — |
case-16 | pass→pass | 4,383 | 2,443 | -44% | 1 | 1 | 0% | 646 | 552 | -15% | 0 | 0 | — |
case-17 | pass→pass | 10,040 | 2,800 | -72% | 1 | 1 | 0% | 1,640 | 624 | -62% | 0 | 0 | — |
case-18 | pass→pass | 11,326 | 3,588 | -68% | 1 | 1 | 0% | 1,729 | 666 | -61% | 0 | 0 | — |
case-19 | pass→pass | 12,112 | 3,425 | -72% | 1 | 1 | 0% | 1,812 | 766 | -58% | 0 | 0 | — |
case-20 | pass→pass | 8,674 | 7,244 | -16% | 1 | 1 | 0% | 1,520 | 1,398 | -8% | 0 | 0 | — |
case-21 | pass→pass | 13,381 | 5,076 | -62% | 1 | 1 | 0% | 2,150 | 949 | -56% | 0 | 0 | — |
case-22 | pass→pass | 13,583 | 9,173 | -32% | 1 | 1 | 0% | 2,427 | 1,783 | -27% | 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. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.