Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →ALWAYS use this when writing docs
.claude/skills/revokslab-docs/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -46% | 0% |
You are an expert technical documentation writer
You are not verbose
Use a relaxed and friendly tone
The title of the page should be a word or a 2-3 word phrase
The description should be one short line, should not start with "The", should avoid repeating the title of the page, should be 5-10 words long
Chunks of text should not be more than 2 sentences long
Each section is separated by a divider of 3 dashes
The section titles are short with only the first letter of the word capitalized
The section titles are in the imperative mood
The section titles should not repeat the term used in the page title, for example, if the page title is "Models", avoid using a section title like "Add new models". This might be unavoidable in some cases, but try to avoid it.
Check out the /packages/web/src/content/docs/docs/index.mdx as an example.
For JS or TS code snippets remove trailing semicolons and any trailing commas that might not be needed.
If you are making a commit prefix the commit message with docs:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 15,757 | 8,103 | -49% | 1 | 1 | 0% | 3,117 | 1,806 | -42% | 0 | 0 | — |
case-08 | fail→pass | 14,018 | 9,485 | -32% | 1 | 1 | 0% | 2,962 | 2,233 | -25% | 0 | 0 | — |
case-02 | fail→pass | 12,958 | 7,471 | -42% | 1 | 1 | 0% | 2,237 | 1,619 | -28% | 0 | 0 | — |
case-03 | fail→fail | 14,361 | 9,596 | -33% | 1 | 1 | 0% | 2,689 | 2,294 | -15% | 0 | 0 | — |
case-04 | fail→pass | 14,708 | 9,158 | -38% | 1 | 1 | 0% | 3,037 | 2,189 | -28% | 0 | 0 | — |
case-05 | fail→pass | 17,054 | 7,137 | -58% | 1 | 1 | 0% | 3,022 | 1,630 | -46% | 0 | 0 | — |
case-06 | fail→pass | 14,834 | 8,230 | -45% | 1 | 1 | 0% | 2,806 | 2,055 | -27% | 0 | 0 | — |
case-07 | fail→pass | 16,515 | 6,591 | -60% | 1 | 1 | 0% | 3,622 | 1,612 | -55% | 0 | 0 | — |
case-09 | fail→pass | 21,720 | 10,162 | -53% | 1 | 1 | 0% | 4,167 | 2,429 | -42% | 0 | 0 | — |
case-10 | fail→pass | 18,235 | 13,580 | -26% | 1 | 1 | 0% | 3,720 | 3,071 | -17% | 0 | 0 | — |
case-11 | fail→pass | 15,454 | 9,707 | -37% | 1 | 1 | 0% | 3,152 | 2,231 | -29% | 0 | 0 | — |
case-12 | fail→pass | 17,737 | 8,277 | -53% | 1 | 1 | 0% | 3,765 | 1,806 | -52% | 0 | 0 | — |
case-13 | fail→pass | 14,815 | 9,485 | -36% | 1 | 1 | 0% | 3,334 | 2,122 | -36% | 0 | 0 | — |
case-14 | fail→pass | 16,321 | 7,165 | -56% | 1 | 1 | 0% | 3,531 | 1,861 | -47% | 0 | 0 | — |
case-15 | fail→pass | 16,463 | 7,447 | -55% | 1 | 1 | 0% | 3,297 | 1,806 | -45% | 0 | 0 | — |
case-16 | fail→pass | 17,131 | 10,707 | -37% | 1 | 1 | 0% | 3,478 | 2,358 | -32% | 0 | 0 | — |
case-17 | pass→pass | 5,970 | 2,764 | -54% | 1 | 1 | 0% | 1,013 | 737 | -27% | 0 | 0 | — |
case-18 | fail→pass | 13,037 | 10,474 | -20% | 1 | 1 | 0% | 2,638 | 2,357 | -11% | 0 | 0 | — |
case-19 | fail→pass | 15,348 | 7,872 | -49% | 1 | 1 | 0% | 3,165 | 1,938 | -39% | 0 | 0 | — |
case-20 | pass→pass | 16,496 | 11,180 | -32% | 1 | 1 | 0% | 3,536 | 2,585 | -27% | 0 | 0 | — |
case-21 | pass→pass | 7,515 | 9,326 | +24% | 1 | 1 | 0% | 1,420 | 1,636 | +15% | 0 | 0 | — |
case-22 | pass→fail | 9,815 | 11,953 | +22% | 1 | 1 | 0% | 2,129 | 2,590 | +22% | 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 +73 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.