Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create a pull request with auto-generated description, issue linking, ROADMAP updates, and PR-metadata validation.
.claude/skills/joshukraine-create-pr/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 70% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 95% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 40% | 0% |
| case-07 | ✓→✗ | ▼ Worse | 46% | 0% |
| case-22 | ✓→✗ | ▼ Worse | 33% | 0% |
Create a pull request from the current branch with proper formatting and issue linking.
All flags are optional. By default, the issue number is inferred from the branch name and the PR is created as ready-to-review.
--issue N: Override automatic issue linking with a specific issue number (or comma-separated list: --issue 123,175)--skip-issue-link: Skip issue linking entirely--draft: Create PR as draftmain/mastergit log origin/<base-branch>..HEAD --oneline (use the base branch identified in step 1)docs/prd/ROADMAP.md[x]) and commit the change to the branch before creating the PR--skip-issue-link: skip this step entirely.--issue N (accepts a comma-separated list, e.g. --issue 123,175) → use directly as Close targets; skip discovery.fix/gh-123-... → P=123). If none can be determined, warn and ask whether to proceed with no linking.gh issue view P --json number,title,body,comments#<number> reference from that output.gh issue view M --json number,title,state — drop anything already closed or that is a PR.closes / fixes / resolves #M, or M is an unchecked task-list item this PR completes → Close (Closes #M)part of / blocked by / relates to / see / parent #M → Reference (Refs #M, no closing keyword)Closes #P.Closes #N per close target, Refs #N per reference.gh pr create with generated title and description--draft flag) unless --draft flag is explicitly passedCloses #N). If either is missing, surface a warning so it can be fixed before review. (These checks previously lived in the retired /review-pr.)markdown## Summary - What this PR does - Key changes made ## Changes - List of changes from commits ## Testing - [ ] Tests pass - [ ] Manual testing done Closes #123
bash# Default: Automatically detects issue from branch name (e.g., fix/gh-123-description) /create-pr # Explicit issue linking (when branch name doesn't contain an issue number) /create-pr --issue 123 # Explicit multi-issue linking (PR closes more than one issue) /create-pr --issue 123,175 # Skip issue linking entirely /create-pr --skip-issue-link
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,673 | 5,790 | -25% | 1 | 1 | 0% | 489 | 1,466 | +200% | 0 | 0 | — |
case-02 | fail→fail | 11,156 | 6,526 | -42% | 1 | 1 | 0% | 1,752 | 1,504 | -14% | 0 | 0 | — |
case-03 | fail→fail | 4,387 | 5,811 | +32% | 1 | 1 | 0% | 426 | 1,447 | +240% | 0 | 0 | — |
case-04 | fail→fail | 7,076 | 4,206 | -41% | 1 | 1 | 0% | 1,263 | 1,956 | +55% | 0 | 0 | — |
case-05 | fail→pass | 8,891 | 6,958 | -22% | 1 | 1 | 0% | 1,454 | 2,471 | +70% | 0 | 0 | — |
case-06 | pass→pass | 6,389 | 2,075 | -68% | 1 | 1 | 0% | 917 | 1,507 | +64% | 0 | 0 | — |
case-07 | pass→fail | 7,665 | 2,492 | -67% | 1 | 1 | 0% | 1,140 | 1,659 | +46% | 0 | 0 | — |
case-08 | fail→fail | 9,548 | 3,420 | -64% | 1 | 1 | 0% | 1,514 | 1,841 | +22% | 0 | 0 | — |
case-09 | pass→pass | 8,821 | 2,223 | -75% | 1 | 1 | 0% | 1,196 | 1,627 | +36% | 0 | 0 | — |
case-10 | pass→pass | 5,483 | 2,952 | -46% | 1 | 1 | 0% | 954 | 1,725 | +81% | 0 | 0 | — |
case-11 | fail→pass | 5,443 | 2,762 | -49% | 1 | 1 | 0% | 872 | 1,697 | +95% | 0 | 0 | — |
case-12 | fail→pass | 8,574 | 4,141 | -52% | 1 | 1 | 0% | 1,360 | 1,903 | +40% | 0 | 0 | — |
case-13 | pass→pass | 4,485 | 2,418 | -46% | 1 | 1 | 0% | 757 | 1,604 | +112% | 0 | 0 | — |
case-14 | pass→pass | 7,213 | 2,884 | -60% | 1 | 1 | 0% | 1,036 | 1,741 | +68% | 0 | 0 | — |
case-15 | pass→pass | 10,197 | 2,616 | -74% | 1 | 1 | 0% | 1,520 | 1,639 | +8% | 0 | 0 | — |
case-16 | pass→pass | 5,806 | 1,791 | -69% | 1 | 1 | 0% | 917 | 1,521 | +66% | 0 | 0 | — |
case-17 | pass→pass | 4,305 | 2,300 | -47% | 1 | 1 | 0% | 627 | 1,547 | +147% | 0 | 0 | — |
case-18 | pass→pass | 9,211 | 5,043 | -45% | 1 | 1 | 0% | 1,577 | 2,158 | +37% | 0 | 0 | — |
case-19 | pass→pass | 8,326 | 1,559 | -81% | 1 | 1 | 0% | 1,395 | 1,467 | +5% | 0 | 0 | — |
case-20 | pass→pass | 8,381 | 3,083 | -63% | 1 | 1 | 0% | 1,554 | 1,615 | +4% | 0 | 0 | — |
case-21 | fail→fail | 8,275 | 7,940 | -4% | 1 | 1 | 0% | 562 | 2,045 | +264% | 0 | 0 | — |
case-22 | pass→fail | 5,843 | 5,431 | -7% | 1 | 1 | 0% | 1,109 | 1,477 | +33% | 0 | 0 | — |
case-23 | fail→fail | 3,847 | 4,821 | +25% | 1 | 1 | 0% | 524 | 1,971 | +276% | 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. 23 cases were attempted, and 19 counted toward the lift figure. The other 4 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 +4 percentage points is the difference between those two pass rates over the 19 comparable cases. 3 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.