Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generate technical documentation, API docs, and content with accessibility and SEO optimization.
.claude/skills/williamzujkowski-content-documentation-generator/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 84% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 88% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 65% | 0% |
Trigger conditions:
Not for:
Time normalization:
NOW_ET using NIST/time.gov semantics (America/New_York, ISO-8601)NOW_ET for all citation access datesInput validation:
content_type must be one of: technical-doc, api-doc, user-guide, tutorialsource must be valid (file exists, spec parseable, or non-empty string)output_format must be: markdown, mdx, or htmlstyle_guide must be: microsoft, google, or write-the-docsSource freshness:
Fast path for 80% of documentation needs:
content_typeOutput: Markdown/MDX document with frontmatter and accessibility-compliant structure.
Decision: If content_type == "api-doc" and OpenAPI spec provided → proceed to T2 for enhanced API documentation. Otherwise STOP at T1.
Extended API documentation from OpenAPI specifications:
Output: Comprehensive API documentation with code samples and interactive elements.
Reserved for:
Content Type Detection:
openapi or swagger key → api-docAbort Conditions:
Accessibility Threshold:
Style Guide Conflicts:
Required fields:
yamldocumentation: type: string description: "Generated documentation content in requested format" format: "markdown | mdx | html" metadata: type: object required: [title, description, keywords, generated_at] properties: title: type: string maxLength: 60 description: type: string maxLength: 160 keywords: type: array items: {type: string} maxItems: 10 generated_at: type: string format: date-time style_guide: type: string enum: [microsoft, google, write-the-docs] accessibility_score: type: integer minimum: 0 maximum: 100 description: "WCAG 2.2 Level AA compliance rating" issues: type: array description: "Accessibility or style issues found (optional)" items: type: object properties: severity: {type: string, enum: [error, warning, info]} message: {type: string} line: {type: integer}
Input:
yamlcontent_type: "api-doc" source: "openapi-spec.yaml" output_format: "markdown" style_guide: "write-the-docs"
Output (abbreviated):
markdown--- title: "User Management API" description: "REST API for user authentication and profile management" keywords: [api, users, authentication, rest] generated_at: "2025-10-25T21:30:36-04:00" style_guide: "write-the-docs" --- # User Management API Manage user accounts and authentication. ## Authentication Bearer token required. Include in `Authorization` header. ## Endpoints ### POST /users/register Create new user account. **Request:**
{"email": "user@example.com", "password": ""}
**Response (201):**{"id": "usr_123", "email": "user@example.com"}
Token budgets (mandatory):
Safety:
Auditability:
Determinism:
Style Guides:
Documentation Tools:
Accessibility:
OpenAPI:
Templates:
resources/templates/ for reusable Markdown templatesresources/schemas/ for JSON schemas| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,109 | 11,704 | -27% | 1 | 1 | 0% | 4,053 | 5,653 | +39% | 0 | 0 | — |
case-02 | fail→pass | 11,608 | 9,870 | -15% | 1 | 1 | 0% | 2,503 | 4,612 | +84% | 0 | 0 | — |
case-03 | fail→pass | 15,787 | 14,146 | -10% | 1 | 1 | 0% | 3,448 | 5,684 | +65% | 0 | 0 | — |
case-04 | pass→fail | 14,747 | 4,425 | -70% | 1 | 1 | 0% | 2,731 | 3,169 | +16% | 0 | 0 | — |
case-05 | pass→fail | 17,340 | 2,346 | -86% | 1 | 1 | 0% | 3,691 | 2,845 | -23% | 0 | 0 | — |
case-06 | pass→fail | 28,223 | 5,983 | -79% | 1 | 1 | 0% | 6,116 | 3,523 | -42% | 0 | 0 | — |
case-07 | fail→fail | 12,094 | 11,630 | -4% | 1 | 1 | 0% | 2,413 | 4,926 | +104% | 0 | 0 | — |
case-08 | fail→pass | 8,696 | 2,066 | -76% | 1 | 1 | 0% | 1,486 | 2,788 | +88% | 0 | 0 | — |
case-09 | fail→fail | 22,679 | 20,403 | -10% | 1 | 1 | 0% | 5,038 | 7,341 | +46% | 0 | 0 | — |
case-10 | pass→pass | 4,410 | 3,686 | -16% | 1 | 1 | 0% | 836 | 3,069 | +267% | 0 | 0 | — |
case-11 | pass→pass | 12,579 | 14,277 | +13% | 1 | 1 | 0% | 2,390 | 5,167 | +116% | 0 | 0 | — |
case-12 | pass→pass | 15,860 | 10,206 | -36% | 1 | 1 | 0% | 3,102 | 4,556 | +47% | 0 | 0 | — |
case-13 | fail→pass | 14,144 | 10,867 | -23% | 1 | 1 | 0% | 2,754 | 4,544 | +65% | 0 | 0 | — |
case-14 | fail→pass | 16,151 | 11,613 | -28% | 1 | 1 | 0% | 2,974 | 4,724 | +59% | 0 | 0 | — |
case-15 | pass→pass | 11,375 | 5,070 | -55% | 1 | 1 | 0% | 2,233 | 3,521 | +58% | 0 | 0 | — |
case-16 | fail→fail | 9,097 | 3,657 | -60% | 1 | 1 | 0% | 2,119 | 3,080 | +45% | 0 | 0 | — |
case-17 | pass→pass | 12,818 | 10,228 | -20% | 1 | 1 | 0% | 2,442 | 4,377 | +79% | 0 | 0 | — |
case-18 | pass→pass | 12,414 | 10,259 | -17% | 1 | 1 | 0% | 2,736 | 4,604 | +68% | 0 | 0 | — |
case-19 | fail→fail | 26,576 | 14,022 | -47% | 1 | 1 | 0% | 6,167 | 5,539 | -10% | 0 | 0 | — |
case-20 | fail→pass | 10,750 | 9,544 | -11% | 1 | 1 | 0% | 2,172 | 4,315 | +99% | 0 | 0 | — |
case-21 | pass→pass | 17,372 | 12,787 | -26% | 1 | 1 | 0% | 3,461 | 5,002 | +45% | 0 | 0 | — |
case-22 | pass→pass | 10,066 | 8,217 | -18% | 1 | 1 | 0% | 1,882 | 4,134 | +120% | 0 | 0 | — |
case-23 | fail→pass | 12,155 | 4,314 | -65% | 1 | 1 | 0% | 2,344 | 3,245 | +38% | 0 | 0 | — |
case-24 | fail→pass | 13,399 | 6,881 | -49% | 1 | 1 | 0% | 2,139 | 3,660 | +71% | 0 | 0 | — |
case-25 | pass→pass | 16,093 | 18,369 | +14% | 1 | 1 | 0% | 3,273 | 6,474 | +98% | 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. The headline lift of +24 percentage points is the difference between those two pass rates over the 25 comparable cases. 3 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.