Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze git diffs for risk scoring, reviewer recommendations, and change classification. Use when preparing a PR, reviewing a large or cross-module change, or before merging to assess risk and pick reviewers.
.claude/skills/ruvnet-diff-analyze/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-15 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -54% | 0% |
Analyze git diffs for risk, complexity, and reviewer assignment.
Before submitting a PR or after making significant changes, analyze the diff to understand risk level, get reviewer recommendations, and classify the type of change.
mcp__plugin_ruflo-core_ruflo__analyze_diff with the diff content for a comprehensive analysismcp__plugin_ruflo-core_ruflo__analyze_diff-risk for a quantified risk assessmentmcp__plugin_ruflo-core_ruflo__analyze_diff-classify to categorize (feature, bugfix, refactor, etc.)mcp__plugin_ruflo-core_ruflo__analyze_diff-reviewers for recommended reviewers based on code ownershipmcp__plugin_ruflo-core_ruflo__analyze_diff-stats for line counts, file counts, complexity metricsmcp__plugin_ruflo-core_ruflo__analyze_file-risk for per-file risk breakdown| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-21 | pass→pass | 8,635 | 6,071 | -30% | 1 | 1 | 0% | 1,693 | 1,398 | -17% | 0 | 0 | — |
case-22 | pass→pass | 7,490 | 5,988 | -20% | 1 | 1 | 0% | 1,354 | 1,365 | +1% | 0 | 0 | — |
case-15 | fail→pass | 7,931 | 2,335 | -71% | 1 | 1 | 0% | 1,303 | 855 | -34% | 0 | 0 | — |
case-01 | fail→fail | 2,879 | 6,061 | +111% | 1 | 1 | 0% | 477 | 831 | +74% | 0 | 0 | — |
case-20 | pass→pass | 9,815 | 5,561 | -43% | 1 | 1 | 0% | 2,061 | 1,396 | -32% | 0 | 0 | — |
case-02 | fail→pass | 5,708 | 3,884 | -32% | 1 | 1 | 0% | 920 | 832 | -10% | 0 | 0 | — |
case-03 | fail→pass | 6,501 | 2,539 | -61% | 1 | 1 | 0% | 1,094 | 757 | -31% | 0 | 0 | — |
case-04 | fail→pass | 12,848 | 2,112 | -84% | 1 | 1 | 0% | 2,226 | 681 | -69% | 0 | 0 | — |
case-05 | fail→pass | 9,225 | 1,876 | -80% | 1 | 1 | 0% | 1,509 | 698 | -54% | 0 | 0 | — |
case-06 | fail→pass | 10,261 | 2,265 | -78% | 1 | 1 | 0% | 1,794 | 744 | -59% | 0 | 0 | — |
case-07 | fail→pass | 11,693 | 4,182 | -64% | 1 | 1 | 0% | 2,007 | 1,010 | -50% | 0 | 0 | — |
case-08 | pass→pass | 11,832 | 5,575 | -53% | 1 | 1 | 0% | 1,970 | 1,268 | -36% | 0 | 0 | — |
case-09 | pass→pass | 11,709 | 6,557 | -44% | 1 | 1 | 0% | 1,838 | 1,318 | -28% | 0 | 0 | — |
case-10 | pass→pass | 10,844 | 4,478 | -59% | 1 | 1 | 0% | 1,788 | 1,075 | -40% | 0 | 0 | — |
case-11 | pass→pass | 14,477 | 9,597 | -34% | 1 | 1 | 0% | 2,378 | 1,994 | -16% | 0 | 0 | — |
case-12 | pass→pass | 10,637 | 6,507 | -39% | 1 | 1 | 0% | 1,831 | 1,411 | -23% | 0 | 0 | — |
case-13 | fail→pass | 9,651 | 3,548 | -63% | 1 | 1 | 0% | 1,798 | 1,087 | -40% | 0 | 0 | — |
case-14 | fail→fail | 9,290 | 2,525 | -73% | 1 | 1 | 0% | 1,583 | 880 | -44% | 0 | 0 | — |
case-16 | fail→fail | 7,083 | 1,724 | -76% | 1 | 1 | 0% | 1,168 | 711 | -39% | 0 | 0 | — |
case-17 | fail→pass | 12,920 | 4,332 | -66% | 1 | 1 | 0% | 2,358 | 1,196 | -49% | 0 | 0 | — |
case-18 | pass→pass | 9,201 | 15,833 | +72% | 1 | 1 | 0% | 1,529 | 1,092 | -29% | 0 | 0 | — |
case-19 | pass→pass | 11,206 | 1,662 | -85% | 1 | 1 | 0% | 1,754 | 548 | -69% | 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 +41 percentage points is the difference between those two pass rates over the 21 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.