Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Alias for pr-writer. Use when users explicitly ask for "create-pr" or reference the legacy skill name. Redirects to the canonical PR writing workflow.
.claude/skills/sickn33-create-pr/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -19% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 21% | 0% |
This skill name is kept for compatibility.
create-pr or refers to the legacy skill name.pr-writer workflow.Use the available pr-writer skill as the canonical workflow for creating and editing pull requests. If the client requires qualified skill names, use the qualifier for the plugin that supplied this skill rather than assuming an external namespace.
If invoked via create-pr, run the same workflow and conventions documented in pr-writer.
User request:
> Create a pull request for the current branch, follow the repository template, and include the validation results.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,570 | 3,041 | -60% | 1 | 1 | 0% | 1,414 | 646 | -54% | 0 | 0 | — |
case-02 | fail→fail | 4,869 | 7,159 | +47% | 1 | 1 | 0% | 872 | 1,419 | +63% | 0 | 0 | — |
case-03 | fail→fail | 7,115 | 3,751 | -47% | 1 | 1 | 0% | 606 | 835 | +38% | 0 | 0 | — |
case-04 | pass→pass | 8,243 | 6,718 | -19% | 1 | 1 | 0% | 1,625 | 1,565 | -4% | 0 | 0 | — |
case-05 | pass→pass | 6,508 | 5,322 | -18% | 1 | 1 | 0% | 1,426 | 1,374 | -4% | 0 | 0 | — |
case-06 | pass→fail | 6,158 | 4,199 | -32% | 1 | 1 | 0% | 1,237 | 1,063 | -14% | 0 | 0 | — |
case-07 | fail→pass | 9,240 | 4,078 | -56% | 1 | 1 | 0% | 1,827 | 906 | -50% | 0 | 0 | — |
case-08 | pass→pass | 4,811 | 3,993 | -17% | 1 | 1 | 0% | 858 | 863 | +1% | 0 | 0 | — |
case-09 | fail→pass | 5,464 | 3,218 | -41% | 1 | 1 | 0% | 987 | 797 | -19% | 0 | 0 | — |
case-10 | fail→pass | 8,644 | 1,481 | -83% | 1 | 1 | 0% | 1,563 | 491 | -69% | 0 | 0 | — |
case-11 | pass→pass | 11,367 | 4,650 | -59% | 1 | 1 | 0% | 1,870 | 941 | -50% | 0 | 0 | — |
case-12 | pass→pass | 10,013 | 7,882 | -21% | 1 | 1 | 0% | 1,693 | 1,627 | -4% | 0 | 0 | — |
case-13 | fail→pass | 7,128 | 3,715 | -48% | 1 | 1 | 0% | 1,115 | 859 | -23% | 0 | 0 | — |
case-14 | pass→pass | 8,198 | 2,927 | -64% | 1 | 1 | 0% | 1,462 | 710 | -51% | 0 | 0 | — |
case-15 | pass→pass | 5,485 | 2,652 | -52% | 1 | 1 | 0% | 940 | 724 | -23% | 0 | 0 | — |
case-16 | fail→pass | 6,118 | 6,354 | +4% | 1 | 1 | 0% | 1,049 | 1,266 | +21% | 0 | 0 | — |
case-17 | fail→pass | 4,769 | 4,529 | -5% | 1 | 1 | 0% | 854 | 954 | +12% | 0 | 0 | — |
case-18 | fail→pass | 6,419 | 4,303 | -33% | 1 | 1 | 0% | 910 | 915 | +1% | 0 | 0 | — |
case-19 | fail→pass | 8,472 | 2,678 | -68% | 1 | 1 | 0% | 1,412 | 633 | -55% | 0 | 0 | — |
case-20 | fail→pass | 6,966 | 1,714 | -75% | 1 | 1 | 0% | 1,176 | 444 | -62% | 0 | 0 | — |
case-21 | fail→pass | 5,035 | 4,086 | -19% | 1 | 1 | 0% | 887 | 909 | +2% | 0 | 0 | — |
case-22 | fail→pass | 6,565 | 5,364 | -18% | 1 | 1 | 0% | 1,185 | 1,141 | -4% | 0 | 0 | — |
case-23 | pass→pass | 4,605 | 3,810 | -17% | 1 | 1 | 0% | 865 | 776 | -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. 23 cases were attempted. The headline lift of +43 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
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.