home / skills / git-fg / thecattoolkit / claudedoc-master

claudedoc-master skill

safe

This skill helps you locate official Claude Code and Anthropic documentation quickly by fuzzy-searching a vetted catalog before web lookup.

npx playbooks add skill git-fg/thecattoolkit --skill claudedoc-master

Review the files below or copy the command above to add this skill to your agents.

Files (3)

SKILL.md

2.3 KB

---
name: claudedoc-master
description: "MUST USE when searching for official Claude Code or Anthropic platform documentation links (MCP, hooks, skills, CLI). Provides fast fuzzy-search CLI for authoritative Claude documentation."
allowed-tools:
  - Read
  - Bash(uv:*)
  - Bash(chmod:*)
  - WebFetch
---

# ClaudeDoc Master - Documentation Link Router

## Overview

This skill provides a fast fuzzy-search CLI (`scripts/claudedoc.py`) for finding authoritative Claude Code and Anthropic platform documentation links. The script searches a curated catalog of official documentation URLs, enabling quick discovery of relevant documentation without web search.

## When to Use

Use this skill when:
- Searching for official Claude Code documentation (skills, commands, agents, hooks, MCP)
- Finding Anthropic platform API documentation
- Verifying documentation URLs before referencing them
- Looking up CLI reference documentation

**Critical Rule**: Always use this skill BEFORE native web search when seeking Claude Code or Anthropic platform documentation. The catalog contains verified, authoritative links.

## Operating Rules

1. **Search First**: When asked about Claude/SDK features, search documentation links using:
   ```bash
   uv run scripts/claudedoc.py "<keywords>"
   ```
   Returns 5 results by default. Use `--limit N` to change the number of results.
   DO NOT use native web search before checking the catalog.

2. **Read Content**: To fetch documentation content from a URL or slug:
   ```bash
   uv run scripts/claudedoc.py --read <url-or-slug>
   ```

## Usage Examples

### Fuzzy Search (default: 5 results)
```bash
uv run scripts/claudedoc.py "hooks"
```

### Fuzzy Search with Custom Limit
```bash
uv run scripts/claudedoc.py "hooks" --limit 10
```

### Read Content by Slug
```bash
uv run scripts/claudedoc.py --read "hooks"
```

### Read Content by URL
```bash
uv run scripts/claudedoc.py --read "https://code.claude.com/docs/en/hooks.md"
```

### Multiple Keywords Search
```bash
uv run scripts/claudedoc.py "skills frontmatter"
```

## Troubleshooting

- **Permission Denied**: If script execution fails, ensure it's executable:
  ```bash
  chmod +x scripts/claudedoc.py
  ```

- **Script Not Found**: Verify the script exists at `scripts/claudedoc.py`

- **Catalog Not Found**: Ensure `catalog.json` exists in the skill root directory

Overview

This skill provides a fast fuzzy-search CLI for finding authoritative Claude Code and Anthropic platform documentation links. It surfaces verified documentation URLs from a curated catalog, so you can locate official pages without a general web search. Use it to quickly confirm references or read specific docs from the command line.

How this skill works

The CLI queries a local catalog of curated, authoritative documentation URLs and returns the closest matches to your keywords using fuzzy search. It can list top results or fetch and display the content of a documentation slug or URL. Commands support a result limit and direct read mode for quick access to full pages.

When to use it

Before performing any general web search for Claude Code or Anthropic platform documentation
When you need official links for skills, hooks, MCP, agents, or CLI references
To verify documentation URLs before citing them in code or docs
When you want fast, terminal-based discovery of Claude/Anthropic docs
To fetch page content by slug or URL from the curated catalog

Best practices

Always run a fuzzy search with concise keywords before asking follow-up questions about SDK features
Use --limit to broaden results if the first set doesn’t include the page you need
Use --read with a slug or URL to inspect the actual document content before implementation
Keep the local catalog (catalog.json) updated from the authoritative source used in your environment
Ensure scripts/claudedoc.py is executable and available in your project scripts directory

Example use cases

Find the official hooks documentation: uv run scripts/claudedoc.py "hooks"
Locate MCP or platform API docs when drafting integration code
Verify the exact CLI or SDK command reference before writing automation scripts
Read a specific doc page by slug: uv run scripts/claudedoc.py --read "hooks"
Increase result set when exploring related topics: uv run scripts/claudedoc.py "skills" --limit 10

FAQ

What if the CLI returns no results?

Try broader or alternate keywords and increase the --limit. Confirm catalog.json exists and is up to date in the skill root.

How do I view the full document content?

Use the --read flag with a slug or a full URL: uv run scripts/claudedoc.py --read "<slug-or-url>". Ensure the script has read permissions.