Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Generates, formats, and validates technical documentation — including docstrings, OpenAPI/Swagger specs, JSDoc annotations, doc portals, and user guides. Use when adding docstrings to functions or classes, creating API documentation, building documentation sites, or writing tutorials and user guides. Invoke for OpenAPI/Swagger specs, JSDoc, doc portals, getting started guides.
.claude/skills/jeffallan-code-documenter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 270% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 228% | 0% |
| case-06 | ✓→✓ | = Same ✓ | 260% | 0% |
| case-07 | ✓→✓ | = Same ✓ | 257% | 0% |
Documentation specialist for inline documentation, API specs, documentation sites, and developer guides.
Applies to any task involving code documentation, API specs, or developer-facing guides. See the reference table below for specific sub-topics.
python -m doctest file.py for doctest blocks; pytest --doctest-modules for module-wide checkstsc --noEmit to confirm typed examples compilenpx @redocly/cli lint openapi.yamlpythondef fetch_user(user_id: int, active_only: bool = True) -> dict: """Fetch a single user record by ID. Args: user_id: Unique identifier for the user. active_only: When True, raise an error for inactive users. Returns: A dict containing user fields (id, name, email, created_at). Raises: ValueError: If user_id is not a positive integer. UserNotFoundError: If no matching user exists. """
pythondef compute_similarity(vec_a: np.ndarray, vec_b: np.ndarray) -> float: """Compute cosine similarity between two vectors. Parameters ---------- vec_a : np.ndarray First input vector, shape (n,). vec_b : np.ndarray Second input vector, shape (n,). Returns ------- float Cosine similarity in the range [-1, 1]. Raises ------ ValueError If vectors have different lengths. """
typescript/** * Fetches a paginated list of products from the catalog. * * @param {string} categoryId - The category to filter by. * @param {number} [page=1] - Page number (1-indexed). * @param {number} [limit=20] - Maximum items per page. * @returns {Promise<ProductPage>} Resolves to a page of product records. * @throws {NotFoundError} If the category does not exist. * * @example * const page = await fetchProducts('electronics', 2, 10); * console.log(page.items); */ async function fetchProducts( categoryId: string, page = 1, limit = 20 ): Promise<ProductPage> { ... }
Load detailed guidance based on context:
| Topic | Reference | Load When | |-------|-----------|-----------| | Python Docstrings | references/python-docstrings.md | Google, NumPy, Sphinx styles | | TypeScript JSDoc | references/typescript-jsdoc.md | JSDoc patterns, TypeScript | | FastAPI/Django API | references/api-docs-fastapi-django.md | Python API documentation | | NestJS/Express API | references/api-docs-nestjs-express.md | Node.js API documentation | | Coverage Reports | references/coverage-reports.md | Generating documentation reports | | Documentation Systems | references/documentation-systems.md | Doc sites, static generators, search, testing | | Interactive API Docs | references/interactive-api-docs.md | OpenAPI 3.1, portals, GraphQL, WebSocket, gRPC, SDKs | | User Guides & Tutorials | references/user-guides-tutorials.md | Getting started, tutorials, troubleshooting, FAQs |
Depending on the task, provide:
Google/NumPy/Sphinx docstrings, JSDoc, OpenAPI 3.0/3.1, AsyncAPI, gRPC/protobuf, FastAPI, Django, NestJS, Express, GraphQL, Docusaurus, MkDocs, VitePress, Swagger UI, Redoc, Stoplight
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 20,794 | 4,325 | -79% | 1 | 1 | 0% | 4,660 | 2,027 | -57% | 0 | 0 | — |
case-02 | fail→fail | 21,656 | 25,302 | +17% | 1 | 1 | 0% | 5,237 | 7,410 | +41% | 0 | 0 | — |
case-03 | fail→fail | 19,218 | 20,962 | +9% | 1 | 1 | 0% | 3,694 | 4,589 | +24% | 0 | 0 | — |
case-04 | pass→pass | 2,787 | 3,384 | +21% | 1 | 1 | 0% | 513 | 1,900 | +270% | 0 | 0 | — |
case-05 | pass→pass | 2,812 | 2,402 | -15% | 1 | 1 | 0% | 495 | 1,625 | +228% | 0 | 0 | — |
case-06 | pass→pass | 3,082 | 2,789 | -10% | 1 | 1 | 0% | 473 | 1,701 | +260% | 0 | 0 | — |
case-07 | pass→pass | 2,540 | 3,137 | +24% | 1 | 1 | 0% | 507 | 1,812 | +257% | 0 | 0 | — |
case-08 | pass→pass | 7,318 | 4,798 | -34% | 1 | 1 | 0% | 1,396 | 2,014 | +44% | 0 | 0 | — |
case-09 | pass→pass | 11,377 | 10,020 | -12% | 1 | 1 | 0% | 2,038 | 2,951 | +45% | 0 | 0 | — |
case-10 | pass→pass | 8,515 | 4,963 | -42% | 1 | 1 | 0% | 1,720 | 2,087 | +21% | 0 | 0 | — |
case-11 | fail→fail | 5,694 | 1,643 | -71% | 1 | 1 | 0% | 946 | 1,468 | +55% | 0 | 0 | — |
case-12 | pass→pass | 7,960 | 4,757 | -40% | 1 | 1 | 0% | 1,398 | 2,113 | +51% | 0 | 0 | — |
case-13 | pass→pass | 8,240 | 2,021 | -75% | 1 | 1 | 0% | 1,384 | 1,566 | +13% | 0 | 0 | — |
case-14 | fail→pass | 7,543 | 3,149 | -58% | 1 | 1 | 0% | 1,229 | 1,754 | +43% | 0 | 0 | — |
case-15 | pass→pass | 11,203 | 2,208 | -80% | 1 | 1 | 0% | 1,852 | 1,647 | -11% | 0 | 0 | — |
case-16 | pass→pass | 12,743 | 2,402 | -81% | 1 | 1 | 0% | 2,157 | 1,678 | -22% | 0 | 0 | — |
case-17 | pass→pass | 10,203 | 5,884 | -42% | 1 | 1 | 0% | 1,994 | 2,337 | +17% | 0 | 0 | — |
case-18 | pass→pass | 13,636 | 11,527 | -15% | 1 | 1 | 0% | 2,452 | 3,361 | +37% | 0 | 0 | — |
case-19 | pass→pass | 5,448 | 1,786 | -67% | 1 | 1 | 0% | 913 | 1,497 | +64% | 0 | 0 | — |
case-20 | fail→fail | 3,806 | 2,880 | -24% | 1 | 1 | 0% | 809 | 1,654 | +104% | 0 | 0 | — |
case-21 | pass→pass | 12,399 | 11,487 | -7% | 1 | 1 | 0% | 2,870 | 4,112 | +43% | 0 | 0 | — |
case-22 | pass→pass | 8,836 | 9,344 | +6% | 1 | 1 | 0% | 1,757 | 2,859 | +63% | 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.
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.