Skip to main content
Raindrop publishes open-source skills that teach any AI coding assistant (Claude Code, Cursor, Devin, Pi, and anything else that supports the skills format) how to work with Raindrop. Instead of wiring up an integration by hand, point your assistant at a skill and let it do the work.
Skills live at github.com/raindrop-ai/skills and are installed with the skills CLI.

​
Available skills

raindrop-setup

Set up, debug, or extend Raindrop observability in a project — the assistant finds your AI call sites, picks the right integration, and instruments them.

raindrop-investigate

Investigate and triage issues in your AI application using Raindrop’s MCP tools.

​
Install

npx skills add raindrop-ai/skills --skill raindrop-setup
npx skills add raindrop-ai/skills --skill raindrop-investigate
Once installed, ask your assistant something like “set up Raindrop in this project” or “add Raindrop tracing to my agent” and it will follow the skill end to end: explore the codebase, propose a plan, install the dependency, instrument the AI call sites, and verify events are flowing.

​
When to use a skill vs. an integration page

  • Use a skill when you want your AI coding assistant to do the integration for you, or when you’re debugging or extending an existing Raindrop setup.
  • Use an integration page when you’re wiring things up yourself and want the exact package, snippet, and configuration for your framework.
The two work together — the raindrop-setup skill matches your stack against the same integrations and uses their reference docs under the hood.