update-docs

Update documentation across the repo after API changes. Finds stale references in docs, examples, docstrings, and fixes them.

npx skills add https://github.com/huggingface/openenv --skill update-docs

/update-docs

Find and fix stale documentation after API changes.

EXECUTE THESE STEPS NOW

When this skill is invoked, you MUST execute these steps immediately.

Step 1: Identify Changed Files

Run:

git diff --name-only main...HEAD -- '*.py'

If no changes are found relative to main (e.g., on main or no upstream), fall back to:

git diff --name-only HEAD~1 -- '*.py'

If that also fails, ask the user which files changed.

Collect the list of changed Python files.

Step 2: Extract API Changes

For each changed .py file, compare old vs new signatures using the same ref that worked in Step 1:

# If Step 1 used main...HEAD:
git diff main...HEAD -- <file>

# If Step 1 fell back to HEAD~1:
git diff HEAD~1 -- <file>

Look for changes to:

  • Function/method signatures (def lines)
  • Class names and init signatures
  • Module-level constants and type aliases
  • Removed or renamed public symbols

Build a list of (old_signature, new_signature) pairs.

If no API changes are found (only internal logic changes), report "No API changes detected — no docs update needed" and stop.

Step 3: Spawn Docs Updater Agent

Use the Task tool to spawn the docs-updater agent. IMPORTANT: the description field MUST contain "docs-updater" so the SubagentStop hook fires correctly.

Task tool:
  subagent_type: general-purpose
  description: "docs-updater propagation"
  prompt: |
    You are a docs-updater agent. Read .claude/agents/docs-updater.md
    for your full instructions.

    Here are the API changes to propagate:

    Changed files: <list>
    API changes:
    - `old` → `new`
    ...

    Search the entire repo for references to the old APIs and update
    them to match the new signatures. Follow the process in
    .claude/agents/docs-updater.md exactly.

Step 4: Review Results

After the agent returns:

  • Review the update report
  • Verify no test files were touched
  • Verify changes are minimal and correct

Report the summary to the user.


When to Use

  • After /implement when API signatures changed
  • Before /pre-submit-pr to ensure docs are fresh
  • When refactoring public APIs

When NOT to Use

  • Internal-only changes (no public API affected)
  • Test-only changes
  • Documentation-only changes (no code changed)

Workflow Integration

/write-tests    →  Red (failing tests)
/implement      →  Green (passing tests)
/update-docs    →  Fix stale docs across repo  ← THIS SKILL
/simplify       →  Refactor (optional)
/pre-submit-pr  →  Validate before PR

More skills from huggingface

sync-models
huggingface
Sync chat-ui's model config with the HuggingFace router — add descriptions for new models, flag reasoning-capable ones, enable artifacts for models with 32B+…
custom-blocks
huggingface
Use when the user has written (or wants to write) a `ModularPipelineBlocks` subclass in a local Python file and needs to package it into a Hub-uploadable…
self-review
huggingface
Use before opening a PR, or whenever asked to self-review a diffusers contribution. Applies the same rubric as the `@claude` CI (checks the diff against…
hf-cloud-sagemaker-production-defaults
huggingface
Create a SageMaker endpoint (real-time or async) with autoscaling, CloudWatch alarms, and tagging enabled by default. Use this skill whenever about to create a…
hf-cloud-serving-image-selection
huggingface
Pick the right serving container for a SageMaker model deployment and find its current image URI. Use this skill whenever about to deploy a model to a…
Hugging Face Cli
huggingface
Execute Hugging Face Hub operations using the `hf` CLI. Use when the user needs to download models/datasets/spaces, upload files to Hub repositories, create repos, manage local cache, or run compute jobs on HF infrastructure. Covers authentication, file transfers, repository creation, cache operations, and cloud compute.
Hugging Face Datasets
huggingface
Create and manage datasets on Hugging Face Hub. Supports initializing repos, defining configs/system prompts, streaming row updates, and SQL-based dataset querying/transformation. Designed to work alongside HF MCP server for comprehensive dataset workflows.
Hugging Face Evaluation
huggingface
Add and manage evaluation results in Hugging Face model cards. Supports extracting eval tables from README content, importing scores from Artificial Analysis API, and running custom model evaluations with vLLM/lighteval. Works with the model-index metadata format.