Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Guide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.
.claude/skills/bilal140202-doc-coauthoring/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 25% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 163% | 0% |
| case-17 | ✓→✗ | ▼ Worse | 0% | 0% |
This skill provides a structured workflow for guiding users through collaborative document creation. Act as an active guide, walking users through three stages: Context Gathering, Refinement & Structure, and Reader Testing.
Trigger conditions:
Initial offer: Offer the user a structured workflow for co-authoring the document. Explain the three stages:
Explain that this approach helps ensure the doc works well when others read it (including when they paste it into Claude). Ask if they want to try this workflow or prefer to work freeform.
If user declines, work freeform. If user accepts, proceed to Stage 1.
Goal: Close the gap between what the user knows and what Claude knows, enabling smart guidance later.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-09 | pass→pass | 9,453 | 5,301 | -44% | 1 | 1 | 0% | 1,409 | 1,697 | +20% | 0 | 0 | — |
case-10 | pass→pass | 5,986 | 5,866 | -2% | 1 | 1 | 0% | 962 | 1,692 | +76% | 0 | 0 | — |
case-01 | fail→pass | 7,974 | 4,447 | -44% | 1 | 1 | 0% | 1,258 | 1,572 | +25% | 0 | 0 | — |
case-02 | fail→fail | 13,428 | 6,139 | -54% | 1 | 1 | 0% | 2,064 | 1,771 | -14% | 0 | 0 | — |
case-03 | fail→fail | 8,958 | 5,066 | -43% | 1 | 1 | 0% | 1,442 | 1,681 | +17% | 0 | 0 | — |
case-04 | pass→fail | 4,726 | 7,133 | +51% | 1 | 1 | 0% | 772 | 2,029 | +163% | 0 | 0 | — |
case-05 | pass→pass | 3,631 | 4,067 | +12% | 1 | 1 | 0% | 596 | 1,472 | +147% | 0 | 0 | — |
case-06 | pass→pass | 4,332 | 2,762 | -36% | 1 | 1 | 0% | 648 | 1,307 | +102% | 0 | 0 | — |
case-07 | fail→pass | 6,791 | 4,564 | -33% | 1 | 1 | 0% | 1,126 | 1,559 | +38% | 0 | 0 | — |
case-08 | fail→pass | 10,021 | 5,861 | -42% | 1 | 1 | 0% | 1,504 | 1,598 | +6% | 0 | 0 | — |
case-11 | pass→pass | 8,528 | 4,795 | -44% | 1 | 1 | 0% | 1,529 | 1,687 | +10% | 0 | 0 | — |
case-12 | pass→pass | 8,775 | 6,310 | -28% | 1 | 1 | 0% | 1,487 | 1,864 | +25% | 0 | 0 | — |
case-13 | pass→pass | 9,144 | 6,302 | -31% | 1 | 1 | 0% | 1,500 | 1,870 | +25% | 0 | 0 | — |
case-14 | pass→pass | 13,564 | 7,651 | -44% | 1 | 1 | 0% | 2,216 | 2,007 | -9% | 0 | 0 | — |
case-15 | pass→pass | 15,080 | 11,839 | -21% | 1 | 1 | 0% | 2,536 | 2,587 | +2% | 0 | 0 | — |
case-16 | pass→pass | 17,287 | 9,542 | -45% | 1 | 1 | 0% | 2,890 | 2,381 | -18% | 0 | 0 | — |
case-17 | pass→fail | 11,084 | 6,217 | -44% | 1 | 1 | 0% | 1,848 | 1,851 | +0% | 0 | 0 | — |
case-18 | fail→fail | 7,388 | 5,654 | -23% | 1 | 1 | 0% | 1,144 | 1,746 | +53% | 0 | 0 | — |
case-19 | pass→pass | 15,844 | 7,491 | -53% | 1 | 1 | 0% | 2,521 | 2,056 | -18% | 0 | 0 | — |
case-20 | fail→fail | 6,776 | 3,928 | -42% | 1 | 1 | 0% | 1,002 | 1,471 | +47% | 0 | 0 | — |
case-21 | pass→pass | 14,448 | 12,404 | -14% | 1 | 1 | 0% | 2,302 | 2,663 | +16% | 0 | 0 | — |
case-22 | fail→fail | 9,139 | 6,486 | -29% | 1 | 1 | 0% | 1,504 | 1,893 | +26% | 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 +5 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 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.