Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Advanced git workflows with branch management, conflict resolution, and PR lifecycle
.claude/skills/ruvnet-git-workflow/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 33% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -46% | 0% |
Advanced git workflow automation for branch management and PR lifecycle.
When managing complex git operations — multi-branch workflows, release branching, conflict resolution, or PR coordination.
mcp__plugin_ruflo-core_ruflo__github_repo_analyze for repository health metricsmcp__plugin_ruflo-core_ruflo__analyze_diff-risk before mergingmcp__plugin_ruflo-core_ruflo__github_pr_manage for PR lifecycle operationsmcp__plugin_ruflo-core_ruflo__github_metrics for merge frequency, review times, etc.bashgit checkout -b feat/my-feature # ... make changes ... # analyze diff before PR # create PR with risk assessment
bashgit checkout -b release/v1.2.0 # cherry-pick fixes # analyze all diffs for risk # merge when risk score is acceptable
bashnpx @claude-flow/cli@latest hooks pre-task --description "git workflow"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,514 | 4,706 | -15% | 1 | 1 | 0% | 962 | 352 | -63% | 0 | 0 | — |
case-20 | pass→pass | 3,501 | 3,592 | +3% | 1 | 1 | 0% | 807 | 1,059 | +31% | 0 | 0 | — |
case-02 | fail→fail | 11,361 | 4,048 | -64% | 1 | 1 | 0% | 1,871 | 971 | -48% | 0 | 0 | — |
case-03 | fail→fail | 11,872 | 4,540 | -62% | 1 | 1 | 0% | 2,084 | 342 | -84% | 0 | 0 | — |
case-04 | fail→fail | 10,676 | 8,425 | -21% | 1 | 1 | 0% | 2,143 | 1,336 | -38% | 0 | 0 | — |
case-05 | fail→pass | 6,413 | 1,998 | -69% | 1 | 1 | 0% | 1,122 | 711 | -37% | 0 | 0 | — |
case-06 | fail→fail | 3,131 | 4,344 | +39% | 1 | 1 | 0% | 375 | 583 | +55% | 0 | 0 | — |
case-07 | fail→fail | 12,622 | 2,196 | -83% | 1 | 1 | 0% | 1,986 | 694 | -65% | 0 | 0 | — |
case-08 | fail→fail | 10,163 | 5,457 | -46% | 1 | 1 | 0% | 1,479 | 811 | -45% | 0 | 0 | — |
case-09 | fail→pass | 2,918 | 1,801 | -38% | 1 | 1 | 0% | 473 | 628 | +33% | 0 | 0 | — |
case-10 | pass→pass | 3,793 | 1,672 | -56% | 1 | 1 | 0% | 684 | 611 | -11% | 0 | 0 | — |
case-11 | fail→pass | 6,255 | 2,739 | -56% | 1 | 1 | 0% | 1,181 | 536 | -55% | 0 | 0 | — |
case-12 | fail→pass | 7,779 | 1,674 | -78% | 1 | 1 | 0% | 1,330 | 542 | -59% | 0 | 0 | — |
case-13 | pass→pass | 9,630 | 1,729 | -82% | 1 | 1 | 0% | 1,596 | 577 | -64% | 0 | 0 | — |
case-14 | pass→fail | 5,749 | 3,313 | -42% | 1 | 1 | 0% | 915 | 654 | -29% | 0 | 0 | — |
case-15 | fail→fail | 10,260 | 3,190 | -69% | 1 | 1 | 0% | 1,693 | 908 | -46% | 0 | 0 | — |
case-16 | pass→pass | 9,458 | 3,142 | -67% | 1 | 1 | 0% | 1,562 | 874 | -44% | 0 | 0 | — |
case-17 | fail→fail | 11,378 | 2,059 | -82% | 1 | 1 | 0% | 2,115 | 657 | -69% | 0 | 0 | — |
case-18 | fail→pass | 10,527 | 3,087 | -71% | 1 | 1 | 0% | 1,709 | 918 | -46% | 0 | 0 | — |
case-19 | pass→pass | 4,870 | 1,365 | -72% | 1 | 1 | 0% | 936 | 527 | -44% | 0 | 0 | — |
case-21 | pass→pass | 6,870 | 4,624 | -33% | 1 | 1 | 0% | 1,322 | 1,156 | -13% | 0 | 0 | — |
case-22 | pass→pass | 8,311 | 7,427 | -11% | 1 | 1 | 0% | 1,609 | 1,774 | +10% | 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 17 counted toward the lift figure. The other 5 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 +18 percentage points is the difference between those two pass rates over the 17 comparable cases. 2 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.