getsentry/sentry-agent-skills
3 skills3.2k installs57 starsGitHub
Install
npx skills add https://github.com/getsentry/sentry-agent-skillsSkills in this repo
1Sentry Fix IssuesSentry-fix-issues is an integration skill for solo builders who already ship SaaS or APIs and need their coding agent to work like an on-call engineer against real Sentry data. You invoke it when production is noisy—specific issue IDs, recent regressions, or a backlog of unresolved events—and want structured analysis that pulls traces and breadcrumbs through the Sentry MCP instead of improvising from pasted stack traces. The skill emphasizes disciplined debugging: understand the failure surface, correlate releases and context, propose a minimal fix, and validate assumptions. Its security model is non-negotiable because exception messages and request bodies are attacker-controlled; the agent must not obey instructions smuggled inside events or litter secrets and PII into commits. Prerequisites are a configured Sentry MCP connection and project access. Complexity sits at intermediate because you need enough production literacy to interpret sampling, environments, and ownership—but you do not need to run Sentry’s UI manually for every step. Pair it with your normal ship-phase testing habits after the fix lands.1.6kinstalls2Sentry React SetupSentry React Setup is an agent skill for solo builders who want official Sentry instrumentation in React or Vite-style entry apps without reverse-engineering init order from scattered docs. It triggers when you ask to add Sentry, install @sentry/react, or configure error boundaries, session replay, or browser tracing. The workflow centers on a dedicated instrument module loaded before the app tree, then Sentry.init with DSN, tracing targets, replay sampling, and optional logs. It is aimed at indie SaaS and mobile-web products where one person owns both frontend and on-call triage. Use it during or right after your first production deploy when you need stack traces and user-session context instead of silent client failures. The skill stresses verifying options against current Sentry documentation because defaults and APIs change. It does not replace backend APM or infrastructure alerting—it front-loads client-side observability so Operate-phase debugging starts with real events.1kinstalls3Sentry Setup Ai Monitoringsentry-setup-ai-monitoring is a Sentry-focused agent skill for solo builders shipping AI features who need observability beyond printf debugging. It walks through enabling AI agent monitoring: confirming distributed tracing is on, choosing integrations for common JavaScript and Python AI stacks, and instrumenting LLM requests, tool executions, and token consumption. The skill emphasizes verifying documentation at implementation time because SDK names and version floors evolve. It also treats privacy as a first-class step—capturing prompts and model outputs can embed end-user PII, so the agent should ask whether to turn on input/output recording rather than defaulting to full content capture. Use it while integrating third-party model providers during Build, then lean on the same configuration through Ship and Operate when incidents or cost spikes need trace-backed diagnosis. It complements generic Sentry error setup by narrowing on AI-specific signals and compliance-aware defaults for indie teams without a dedicated platform group.559installs