Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyzes all commits on the current branch vs the base branch, generates a PR title and description, pushes the branch, and creates the PR using `gh`. Includes a summary of changes, test plan, and links to related issues.
.claude/skills/miosa-osa-create-pr/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-22 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-06 | ✓→✗ | ▼ Worse | 3% | 0% |
| case-12 | ✓→✗ | ▼ Worse | -68% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 24% | 0% |
> Create a pull request with summary, test plan, and proper labeling.
/create-pr [--base <branch>] [--draft] [--title "title"]Analyzes all commits on the current branch vs the base branch, generates a PR title and description, pushes the branch, and creates the PR using gh. Includes a summary of changes, test plan, and links to related issues.
git log and git diff against base branch.git push -u origin <branch>.gh pr create with title, body, labels, reviewers.bash# Create PR against main /create-pr # Create draft PR against specific base /create-pr --base develop --draft # Create PR with explicit title /create-pr --title "Add user authentication flow"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→fail | 40,996 | 39,275 | -4% | 1 | 1 | 0% | 1,625 | 1,666 | +3% | 0 | 0 | — |
case-05 | fail→fail | 3,329 | 2,855 | -14% | 1 | 1 | 0% | 383 | 561 | +46% | 0 | 0 | — |
case-01 | fail→fail | 3,982 | 33,122 | +732% | 1 | 1 | 0% | 521 | 679 | +30% | 0 | 0 | — |
case-02 | fail→fail | 40,032 | 3,746 | -91% | 1 | 1 | 0% | 1,695 | 728 | -57% | 0 | 0 | — |
case-03 | fail→fail | 11,381 | 6,646 | -42% | 1 | 1 | 0% | 1,838 | 644 | -65% | 0 | 0 | — |
case-04 | pass→pass | 34,930 | 4,825 | -86% | 1 | 1 | 0% | 844 | 1,049 | +24% | 0 | 0 | — |
case-07 | fail→pass | 12,951 | 2,971 | -77% | 1 | 1 | 0% | 1,352 | 690 | -49% | 0 | 0 | — |
case-08 | pass→pass | 5,437 | 1,417 | -74% | 1 | 1 | 0% | 819 | 524 | -36% | 0 | 0 | — |
case-09 | pass→pass | 6,048 | 1,988 | -67% | 1 | 1 | 0% | 841 | 613 | -27% | 0 | 0 | — |
case-10 | pass→pass | 6,747 | 2,417 | -64% | 1 | 1 | 0% | 1,204 | 569 | -53% | 0 | 0 | — |
case-11 | pass→pass | 10,894 | 7,462 | -32% | 1 | 1 | 0% | 1,863 | 1,582 | -15% | 0 | 0 | — |
case-12 | pass→fail | 9,122 | 1,643 | -82% | 1 | 1 | 0% | 1,585 | 507 | -68% | 0 | 0 | — |
case-13 | pass→pass | 13,003 | 10,123 | -22% | 1 | 1 | 0% | 2,363 | 1,726 | -27% | 0 | 0 | — |
case-14 | pass→pass | 15,132 | 12,041 | -20% | 1 | 1 | 0% | 2,206 | 2,158 | -2% | 0 | 0 | — |
case-15 | pass→pass | 4,427 | 2,171 | -51% | 1 | 1 | 0% | 820 | 653 | -20% | 0 | 0 | — |
case-16 | pass→pass | 6,393 | 2,298 | -64% | 1 | 1 | 0% | 861 | 566 | -34% | 0 | 0 | — |
case-17 | pass→pass | 11,485 | 7,235 | -37% | 1 | 1 | 0% | 2,042 | 1,671 | -18% | 0 | 0 | — |
case-18 | pass→pass | 3,624 | 3,341 | -8% | 1 | 1 | 0% | 604 | 712 | +18% | 0 | 0 | — |
case-19 | pass→pass | 6,693 | 2,763 | -59% | 1 | 1 | 0% | 1,076 | 622 | -42% | 0 | 0 | — |
case-20 | pass→pass | 9,469 | 1,773 | -81% | 1 | 1 | 0% | 1,460 | 504 | -65% | 0 | 0 | — |
case-21 | pass→pass | 3,127 | 2,001 | -36% | 1 | 1 | 0% | 501 | 504 | +1% | 0 | 0 | — |
case-22 | fail→pass | 10,274 | 7,320 | -29% | 1 | 1 | 0% | 1,894 | 1,690 | -11% | 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 0 percentage points is the difference between those two pass rates over the 22 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.