getsentry sentry-for-cursor
OFFICIALLABSCO SUMMARY
Five of the six set up Sentry instrumentation in a project. sentry-ios-swift-setup installs the sentry-cocoa SDK for iOS and SwiftUI apps; sentry-setup-ai-monitoring detects an installed OpenAI, Anthropic, Vercel AI, LangChain, or Google GenAI SDK and wires up Sentry's AI agent monitoring for it; sentry-setup-logging, sentry-setup-metrics, and sentry-setup-tracing each configure the matching Sentry feature — structured logs, counters/gauges/distributions, and performance spans and transactions — across JavaScript, TypeScript, Python, Ruby, React, Next.js, and Node.js, with some language coverage differing skill to skill. The sixth, sentry-code-review, does something different: it reads Sentry's own comments on a GitHub pull request and resolves the issues Sentry flagged, either for a PR you name or by finding recent PRs with Sentry feedback on its own.
Five of the six carry an account_key dependency in our data — they need a Sentry auth token or project DSN on file before they do anything — and the sixth needs a local tool instead, which fits sentry-code-review working against a GitHub PR rather than writing SDK config into a project. This was built for a team already on Sentry, working in Cursor, that wants an agent to wire up instrumentation or close out Sentry-flagged PR feedback without hand-writing the setup code itself.
READ THE FULL ANALYSIS
Nothing here should be installed from this repository going forward. Its README has no other content — it is a migration notice telling you to uninstall this plugin, search the Cursor marketplace for "sentry" instead, or clone getsentry/sentry-for-ai directly, which it says serves both Cursor and Claude Code from one repository and carries over every skill, command, and MCP configuration this one had. We describe the six skills above because their names and descriptions are still on record, not as a recommendation to use this repository.
Checked 20 September 2026 against the repository's own archival notice and its six recorded skills; genuinely Sentry's own repository, MIT-licensed, and we have not run any of the six ourselves.
WHAT'S INSIDE
6 showing · 6 totalsentry-code-review
Reads the bug comments Sentry's bot leaves on a GitHub pull request, fixes the issues it flagged, and reports what got fixed versus what still needs a person.
sentry-ios-swift-setup
When an iPhone app crashes on somebody else's phone, nobody tells you. This is the wiring that makes a Swift app report back for itself, put in place from its very first line of code onward.
sentry-setup-ai-monitoring
AI features fail quietly: a model call is slow, a chain of steps goes wrong somewhere in the middle, and the cost creeps up out of sight. This works out which AI library the project already uses and switches on the matching recording, so each call leaves a trail you can open up afterwards.
sentry-setup-logging
A program's own running commentary — the lines that normally scroll past in a terminal window and are gone — kept in Sentry instead, where it can be searched later and lined up against the moment something broke.
sentry-setup-metrics
Errors tell you something broke; this is for the numbers that tell you whether things are going well at all — taken from inside the running code and plotted over time in Sentry.
sentry-setup-tracing
A slow page is easy to notice and hard to explain. This puts a stopwatch inside the code itself, so instead of guessing where the time went, you can look.
HOW TO GET IT
npx skills add getsentry/sentry-for-cursornpx skills add getsentry/sentry-for-cursor --skill <name> --full-depthPick the skill name from the Skills tab — each entry there installs independently.