Skip to content
cccskills
FreePublished on GitHub

update-api-docs

Generate or update EmbodiChain Sphinx API documentation for public Python exports. Use when the API docs checker or CI reports missing __all__ exports, after adding or changing public APIs, or when asked to fill, generate, or synchronize API-reference pages and their descriptions.

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236 stars(GitHub 236 + cccskills 0)
Works withX (Twitter)

Run score Not tested

See how to install

Files included (2)

  • SKILL.md3.4 KB
  • agents/openai.yaml228 B

SKILL.md (original)

Review the instructions the agent will receive before you install.

Update API Docs

Generate useful API documentation for every public export reported by the read-only checker. Treat static __all__ declarations as the public API contract and preserve existing hand-written documentation.

Workflow

  1. Run the checker in machine-readable mode:

    python docs/scripts/check_api_docs.py --format json
    

    Exit status 1 is expected when exports are missing. Read the JSON report; if missing_count is zero, report that the docs are aligned and stop.

  2. Group missing entries by module. Read each reported source file and inspect the exported definition, signature, type annotations, and docstring. Locate existing API pages with:

    rg -n "automodule:: <module>|currentmodule:: <module>" docs/source/api_reference
    
  3. Choose the documentation location:

    • Add the export to an existing curated module page when one exists. Follow that page's headings, autosummary groups, and detailed autodoc directives.
    • Document a package-level re-export under its public import path, not only under the implementation module.
    • If no suitable curated page exists, add or extend the module section in docs/source/api_reference/public_api.rst. This file is an agent-maintained fallback, not checker output. Keep fallback module headings sorted by import path and entries in their declared __all__ order to minimize diff churn.
  4. Write documentation that explains the API:

    • Add the export to the appropriate autosummary block.
    • Add the matching autoclass, autofunction, autodata, or other detailed directive when the surrounding page provides detailed entries.
    • Add a concise section overview when names alone do not explain the group.
    • If the source docstring is missing or too vague for autodoc, improve it with a meaningful summary and Google-style Args, Returns, and Raises sections where applicable.
    • Derive descriptions from the implementation and tests. Do not invent behavior, examples, guarantees, or parameter semantics.
  5. Keep the change scoped to documentation. Do not alter runtime behavior, signatures, or __all__ merely to silence the checker. Do not replace curated prose with generic generated text or add placeholders such as "part of the public API."

  6. Rerun the checker until it reports zero missing exports:

    python docs/scripts/check_api_docs.py
    
  7. Run Black on every changed Python file, then validate the documentation workflow:

    pytest tests/docs/test_check_api_docs.py -q --confcutdir=tests/docs
    python -m sphinx -b dummy docs/source docs/build/api-docs-check
    

    When source docstrings were changed, also run focused tests for those modules. Fix new Sphinx warnings caused by the edit; distinguish them from unrelated pre-existing warnings.

Completion Report

Report the documented import paths, the pages or docstrings updated, and the validation results. If an export cannot be documented accurately from the repository, identify the exact missing semantic information instead of guessing.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Add a new simulation atomic action or motion primitive to EmbodiChain's typed planning and execution framework. Use when implementing a new skill, goal contract, action planner, symbolic effect, registration entry, documentation, and tests for AtomicActionEngine.

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236Updated Oct 26, 2026

Claude adapter for the canonical EmbodiChain add-atomic-action skill.

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236Updated Oct 26, 2026

Claude adapter for the canonical EmbodiChain add-embodiment-component skill.

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236Updated Oct 26, 2026

Add or update a reusable EmbodiChain embodiment YAML component that owns one simulation robot, its sensor suite, and an optional Task Program skill_profile. Use for configs/components/embodiments, new Task Program-compatible robots, or multi-embodiment task deployment.

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236Updated Oct 26, 2026

Claude adapter for the canonical EmbodiChain add-functor skill.

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236Updated Oct 26, 2026

Use when adding a new observation, event, reward, action, dataset, or randomization functor to an EmbodiChain environment

The English summary is being prepared. Showing the original description.

DexForce/EmbodiChain236Updated Oct 26, 2026

See all skills by DexForce

Report a problem with this skill