Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Write, review, and edit documentation files with consistent structure, tone, and technical accuracy. Use when creating docs, reviewing markdown files, writing READMEs, updating `/docs` directories, or when user says "write documentation", "review this doc", "improve this README", "create a guide", or "edit markdown". Do NOT use for code comments, inline JSDoc, or API reference generation.
.claude/skills/tech-leads-club-docs-writer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 189% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -30% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -1% | 0% |
docs-writer skill instructionsAs an expert technical writer and editor, your goal is to produce and refine documentation that is accurate, clear, consistent, and easy for users to understand. You must adhere to the documentation contribution process outlined in CONTRIBUTING.md.
packages/ directory, to ensure your work is backed by the implementation and to identify any gaps.docs/ directory that need to be modified. Always read the latest version of a file before you begin work.docs/sidebar.json needs to be updated. Make sure all links are up to date.references/style-guide.md. Read this file to understand the project's documentation standards.replace and write_file: Use file system tools to apply your planned changes. For small edits, replace is preferred. For new files or large rewrites, write_file is more appropriate.npm run format| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 53,298 | 3,468 | -93% | 1 | 1 | 0% | 3,052 | 1,160 | -62% | 0 | 0 | — |
case-02 | fail→fail | 3,161 | 2,744 | -13% | 1 | 1 | 0% | 483 | 1,055 | +118% | 0 | 0 | — |
case-03 | fail→fail | 2,640 | 2,414 | -9% | 1 | 1 | 0% | 218 | 965 | +343% | 0 | 0 | — |
case-04 | fail→pass | 2,968 | 2,945 | -1% | 1 | 1 | 0% | 384 | 1,110 | +189% | 0 | 0 | — |
case-05 | fail→pass | 12,603 | 4,092 | -68% | 1 | 1 | 0% | 1,963 | 1,378 | -30% | 0 | 0 | — |
case-06 | pass→pass | 3,959 | 2,425 | -39% | 1 | 1 | 0% | 623 | 1,024 | +64% | 0 | 0 | — |
case-07 | fail→pass | 4,236 | 1,801 | -57% | 1 | 1 | 0% | 606 | 921 | +52% | 0 | 0 | — |
case-08 | fail→pass | 6,561 | 1,543 | -76% | 1 | 1 | 0% | 1,065 | 842 | -21% | 0 | 0 | — |
case-09 | fail→pass | 4,726 | 1,809 | -62% | 1 | 1 | 0% | 959 | 945 | -1% | 0 | 0 | — |
case-10 | pass→pass | 5,014 | 1,240 | -75% | 1 | 1 | 0% | 784 | 835 | +7% | 0 | 0 | — |
case-11 | pass→pass | 9,729 | 64,945 | +568% | 1 | 1 | 0% | 1,455 | 1,423 | -2% | 0 | 0 | — |
case-12 | pass→pass | 9,656 | 3,568 | -63% | 1 | 1 | 0% | 1,663 | 1,293 | -22% | 0 | 0 | — |
case-13 | pass→pass | 5,305 | 2,005 | -62% | 1 | 1 | 0% | 808 | 1,004 | +24% | 0 | 0 | — |
case-14 | pass→pass | 8,106 | 2,608 | -68% | 1 | 1 | 0% | 1,392 | 1,112 | -20% | 0 | 0 | — |
case-15 | pass→pass | 3,372 | 1,417 | -58% | 1 | 1 | 0% | 566 | 839 | +48% | 0 | 0 | — |
case-16 | pass→pass | 10,000 | 3,090 | -69% | 1 | 1 | 0% | 1,547 | 1,163 | -25% | 0 | 0 | — |
case-17 | pass→pass | 11,166 | 5,935 | -47% | 1 | 1 | 0% | 1,853 | 1,533 | -17% | 0 | 0 | — |
case-18 | fail→pass | 51,199 | 3,335 | -93% | 1 | 1 | 0% | 1,092 | 1,109 | +2% | 0 | 0 | — |
case-19 | pass→pass | 6,820 | 6,241 | -8% | 1 | 1 | 0% | 1,061 | 1,641 | +55% | 0 | 0 | — |
case-20 | pass→pass | 3,169 | 5,893 | +86% | 1 | 1 | 0% | 456 | 1,637 | +259% | 0 | 0 | — |
case-21 | pass→pass | 4,322 | 5,898 | +36% | 1 | 1 | 0% | 864 | 1,685 | +95% | 0 | 0 | — |
case-22 | pass→pass | 6,543 | 3,969 | -39% | 1 | 1 | 0% | 1,156 | 1,326 | +15% | 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 +27 percentage points is the difference between those two pass rates over the 22 comparable cases.
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.