A Claude Code skill for working with the Catalyzed data platform API.
This skill provides Claude Code with knowledge about the Catalyzed API, enabling it to help you:
- Create and manage datasets and tables
- Ingest and query data using SQL
- Build and run data pipelines
- Set up knowledge bases for vector search
- Authenticate and manage API tokens
When invoked, the skill fetches live documentation from docs.catalyzed.ai and falls back to bundled reference material if offline.
git clone https://github.com/moderately-ai/catalyzed-skill ~/.claude/skills/catalyzed# From your project root
git clone https://github.com/moderately-ai/catalyzed-skill .claude/skills/catalyzed
# Either add to .gitignore or commit to share with your team
echo ".claude/skills/" >> .gitignore # To ignore
# OR
git add .claude/skills/catalyzed # To shareIn any Claude Code session:
/catalyzed
Then ask questions about the Catalyzed API, request code examples, or get help with integration.
After invoking /catalyzed:
- "How do I create a dataset and table?"
- "Show me how to insert data into a table"
- "Write a Python script to query my events table"
- "How do I set up a knowledge base for semantic search?"
- "What's the authentication flow for API tokens?"
cd ~/.claude/skills/catalyzed # or .claude/skills/catalyzed
git pull| File | Description |
|---|---|
SKILL.md |
Main skill definition with dynamic docs loading |
api-reference.md |
Complete API endpoint reference |
examples.md |
Code examples in curl, TypeScript, and Python |
README.md |
This file |
- When you invoke
/catalyzed, Claude Code loadsSKILL.md - The skill attempts to fetch live documentation from
https://docs.catalyzed.ai/llms.txt - If successful, Claude has access to the latest API documentation
- If offline, Claude uses the bundled fallback content in
SKILL.md - For detailed topics, Claude can use
WebFetchto retrieve specific documentation pages
MIT