Skip to content

Home

Jump to bottom
Daniel Griesser edited this page Mar 9, 2026 · 4 revisions

What is sentry-for-ai?

sentry-for-ai is a repo that gives AI coding agents expert-level knowledge of Sentry. It works across Claude Code, Cursor, and GitHub Copilot since most things are Skills based.

The idea: When a developer asks their AI assistant to "add Sentry to my Next.js app" or "set up error monitoring", the agent shouldn't have to guess or rely on stale training data. It should have access to current, verified, opinionated setup guides — the same way a senior engineer who's done it 50 times would walk you through it.

How it works: The repo ships SDK skills — one per platform (Go, Python, React, Next.js, Flutter, etc.). Each skill is a multi-file bundle containing a setup wizard and deep-dive references for every Sentry feature the SDK supports: error monitoring, tracing, profiling, logging, session replay, and more.

When a developer invokes a skill, the agent:

  1. Scans their project to detect the framework, existing dependencies, and whether Sentry is already installed
  2. Recommends which Sentry features to enable based on what it found — opinionated, not open-ended
  3. Walks through installation, configuration, and verification step by step
  4. Suggests companion SDKs if it detects a frontend/backend pair

The skills are designed to be consumed by any AI coding tool that supports the Agent Skills spec or MCP. The plugin also connects to the Sentry MCP server so agents can query your Sentry environment directly — look up issues, check error rates, investigate performance.

Why this matters: Sentry SDK setup is deceptively complex. There are dozens of config options, framework-specific middleware, source map upload pipelines, and feature flags that change across versions. Getting it wrong means silent failures — events that never arrive, stack traces that show minified code, performance data that's missing. Skills encode the right way to do it so every developer gets a production-grade setup on the first try, regardless of which AI tool they use.

When should you use the skills?: For really any SDK (Sentry) related tasks. While the skills are optimized for helping people onboard, they are equally as effective with helping updating/improving peoples instrumentation.

Links

Clone this wiki locally