Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Review pull requests for code quality, security issues, and best practices. Use when reviewing PRs, checking code changes, or analyzing diffs before merge.
.claude/skills/onewave-ai-git-pr-reviewer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-13 | ✗→✓ | ▲ Improved | 126% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-08 | ✓→✗ | ▼ Worse | -59% | 0% |
When reviewing a pull request:
git diff main...HEAD or git diff <base-branch>...HEADmarkdown## PR Review Summary ### Overview [Brief summary of changes] ### Issues Found #### Critical - [Issue description + file:line] #### Warnings - [Issue description + file:line] #### Suggestions - [Improvement ideas] ### Approval Status [APPROVE / REQUEST CHANGES / NEEDS DISCUSSION]
bash# Review current branch against main git diff main...HEAD --stat git diff main...HEAD
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,360 | 25,960 | +673% | 1 | 1 | 0% | 449 | 1,325 | +195% | 0 | 0 | — |
case-02 | fail→fail | 13,943 | 3,897 | -72% | 1 | 1 | 0% | 2,209 | 647 | -71% | 0 | 0 | — |
case-03 | fail→fail | 4,719 | 2,263 | -52% | 1 | 1 | 0% | 744 | 628 | -16% | 0 | 0 | — |
case-13 | fail→pass | 9,050 | 15,763 | +74% | 1 | 1 | 0% | 1,377 | 3,110 | +126% | 0 | 0 | — |
case-04 | pass→pass | 5,641 | 5,209 | -8% | 1 | 1 | 0% | 938 | 1,179 | +26% | 0 | 0 | — |
case-05 | pass→pass | 9,334 | 5,202 | -44% | 1 | 1 | 0% | 1,569 | 1,231 | -22% | 0 | 0 | — |
case-06 | pass→pass | 11,889 | 9,850 | -17% | 1 | 1 | 0% | 1,889 | 1,803 | -5% | 0 | 0 | — |
case-07 | fail→fail | 16,582 | 23,391 | +41% | 1 | 1 | 0% | 2,028 | 2,220 | +9% | 0 | 0 | — |
case-14 | fail→fail | 18,626 | 2,449 | -87% | 1 | 1 | 0% | 2,460 | 698 | -72% | 0 | 0 | — |
case-08 | pass→fail | 10,258 | 2,846 | -72% | 1 | 1 | 0% | 1,854 | 762 | -59% | 0 | 0 | — |
case-09 | fail→pass | 9,602 | 6,959 | -28% | 1 | 1 | 0% | 1,654 | 1,580 | -4% | 0 | 0 | — |
case-10 | fail→pass | 9,860 | 14,528 | +47% | 1 | 1 | 0% | 1,571 | 1,617 | +3% | 0 | 0 | — |
case-11 | fail→fail | 12,120 | 5,322 | -56% | 1 | 1 | 0% | 1,924 | 656 | -66% | 0 | 0 | — |
case-12 | pass→pass | 9,178 | 8,372 | -9% | 1 | 1 | 0% | 1,622 | 1,786 | +10% | 0 | 0 | — |
case-15 | pass→pass | 7,603 | 4,510 | -41% | 1 | 1 | 0% | 1,160 | 869 | -25% | 0 | 0 | — |
case-16 | fail→pass | 6,196 | 4,784 | -23% | 1 | 1 | 0% | 802 | 906 | +13% | 0 | 0 | — |
case-17 | fail→fail | 19,108 | 5,234 | -73% | 1 | 1 | 0% | 2,720 | 626 | -77% | 0 | 0 | — |
case-18 | fail→fail | 2,894 | 2,909 | +1% | 1 | 1 | 0% | 439 | 654 | +49% | 0 | 0 | — |
case-19 | pass→pass | 7,578 | 5,212 | -31% | 1 | 1 | 0% | 1,285 | 1,076 | -16% | 0 | 0 | — |
case-20 | pass→pass | 15,471 | 11,045 | -29% | 1 | 1 | 0% | 2,439 | 2,131 | -13% | 0 | 0 | — |
case-21 | pass→fail | 13,400 | 2,274 | -83% | 1 | 1 | 0% | 2,245 | 739 | -67% | 0 | 0 | — |
case-22 | fail→fail | 16,288 | 10,217 | -37% | 1 | 1 | 0% | 2,632 | 2,084 | -21% | 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 21 counted toward the lift figure. The other 1 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 0 percentage points is the difference between those two pass rates over the 21 comparable cases. 5 cases got worse with the skill loaded, and they are 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.