Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →9 editing & proofreading skills. Trigger: polishing drafts, academic tone, proofreading, translation. Design: style checkers and editing workflows for clear, concise academic English.
.claude/skills/brycewang-stanford-polish-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -69% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -46% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | academic-tone-guide | Adjust writing tone and register for academic audiences and venues | | academic-translation-guide | Academic translation, post-editing, and Chinglish correction guide | | academic-writing-refiner | Checklist-driven academic English polishing and Chinglish correction | | ai-writing-humanizer | Remove AI-generated patterns to produce natural, authentic academic writing | | chinese-text-humanizer | Transform AI-generated Chinese text into natural academic writing style | | conciseness-editing-guide | Eliminate wordiness and redundancy in academic prose for clarity | | grammar-checker-guide | Use grammar and style checking tools to polish academic manuscripts | | paper-polish-guide | Review and polish LaTeX research papers for clarity and style | | plagiarism-detection-guide | Use plagiarism detection tools and ensure manuscript originality |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 14,813 | 16,607 | +12% | 1 | 1 | 0% | 2,358 | 998 | -58% | 0 | 0 | — |
case-02 | fail→fail | 8,378 | 4,825 | -42% | 1 | 1 | 0% | 1,334 | 830 | -38% | 0 | 0 | — |
case-03 | pass→fail | 11,032 | 6,868 | -38% | 1 | 1 | 0% | 1,789 | 818 | -54% | 0 | 0 | — |
case-04 | fail→pass | 13,581 | 3,148 | -77% | 1 | 1 | 0% | 2,175 | 1,011 | -54% | 0 | 0 | — |
case-05 | fail→pass | 10,605 | 3,046 | -71% | 1 | 1 | 0% | 1,602 | 916 | -43% | 0 | 0 | — |
case-15 | fail→fail | 18,163 | 7,701 | -58% | 1 | 1 | 0% | 2,826 | 1,042 | -63% | 0 | 0 | — |
case-06 | fail→pass | 16,981 | 2,508 | -85% | 1 | 1 | 0% | 2,704 | 828 | -69% | 0 | 0 | — |
case-07 | fail→fail | 7,390 | 3,249 | -56% | 1 | 1 | 0% | 1,076 | 932 | -13% | 0 | 0 | — |
case-08 | fail→fail | 13,911 | 2,702 | -81% | 1 | 1 | 0% | 2,001 | 798 | -60% | 0 | 0 | — |
case-09 | fail→pass | 12,135 | 3,934 | -68% | 1 | 1 | 0% | 2,054 | 1,110 | -46% | 0 | 0 | — |
case-10 | fail→fail | 11,574 | 5,383 | -53% | 1 | 1 | 0% | 1,936 | 947 | -51% | 0 | 0 | — |
case-11 | fail→fail | 22,516 | 4,664 | -79% | 1 | 1 | 0% | 1,775 | 695 | -61% | 0 | 0 | — |
case-12 | fail→fail | 10,757 | 17,371 | +61% | 1 | 1 | 0% | 1,638 | 2,480 | +51% | 0 | 0 | — |
case-13 | fail→pass | 9,833 | 5,140 | -48% | 1 | 1 | 0% | 1,641 | 1,276 | -22% | 0 | 0 | — |
case-14 | fail→pass | 13,654 | 2,891 | -79% | 1 | 1 | 0% | 2,034 | 831 | -59% | 0 | 0 | — |
case-16 | fail→pass | 13,491 | 3,336 | -75% | 1 | 1 | 0% | 1,921 | 948 | -51% | 0 | 0 | — |
case-17 | fail→pass | 17,012 | 6,188 | -64% | 1 | 1 | 0% | 2,765 | 863 | -69% | 0 | 0 | — |
case-18 | pass→pass | 6,336 | 2,797 | -56% | 1 | 1 | 0% | 981 | 904 | -8% | 0 | 0 | — |
case-19 | fail→pass | 8,055 | 2,948 | -63% | 1 | 1 | 0% | 1,196 | 889 | -26% | 0 | 0 | — |
case-20 | pass→pass | 16,433 | 2,869 | -83% | 1 | 1 | 0% | 2,396 | 865 | -64% | 0 | 0 | — |
case-21 | fail→fail | 14,379 | 2,418 | -83% | 1 | 1 | 0% | 2,604 | 789 | -70% | 0 | 0 | — |
case-22 | fail→pass | 8,513 | 2,325 | -73% | 1 | 1 | 0% | 1,225 | 733 | -40% | 0 | 0 | — |
case-23 | fail→pass | 6,159 | 1,846 | -70% | 1 | 1 | 0% | 890 | 667 | -25% | 0 | 0 | — |
case-24 | fail→pass | 9,689 | 1,696 | -82% | 1 | 1 | 0% | 1,439 | 576 | -60% | 0 | 0 | — |
case-25 | fail→pass | 10,766 | 5,781 | -46% | 1 | 1 | 0% | 1,820 | 868 | -52% | 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. 25 cases were attempted, and 22 counted toward the lift figure. The other 3 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 +52 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.