This sets up Sentry's AI monitoring so you can track LLM calls, token usage, and agent execution across OpenAI, Anthropic, Vercel AI, LangChain, and others. It detects what AI SDKs you've already installed, configures the right integrations, and handles the tricky part about sampling rates (AI traces need to be sampled as complete trees or you lose visibility). The skill is opinionated about PII: it won't enable prompt and output recording without asking you first, which is smart since you're basically capturing user conversations. Works for both JavaScript and Python, with auto-instrumentation where available and manual span setup when you need it. Good if you want observability without reading through Sentry's docs to figure out which integration matches your stack.
npx skills add https://github.com/getsentry/sentry-for-ai --skill sentry-setup-ai-monitoring