Back to templates
🔭

Sentry Monitoring with autofix

by @plb

Required Toolkits

SentrySlack

Description

A fully automated error-to-fix pipeline. Every morning, this agent triages your Sentry errors, posts a summary to Slack, opens GitHub issues for critical problems, and tags an AI coding agent to investigate and open pull requests — so bugs get fixed before your team even starts their day. What it does:

  • Triage — Pulls unresolved errors from the last 24h across your Sentry projects, filters out noise, and surfaces what actually matters: high-priority unhandled exceptions, high-frequency errors, infrastructure failures, and zombie issues that keep recurring
  • Report — Posts a concise summary to Slack with a detailed breakdown in a thread, so your team has full context
  • Create issues — Opens GitHub issues for every critical problem with full error details, occurrence data, and impact analysis
  • Trigger fixes — Tags an AI coding agent (@twill or @claude) on each issue to automatically investigate, suggest a fix, and open a pull request
  • Stay silent — Does nothing when everything is fine. No "all clear" spam.

You need to setup a GITHUB_TOKEN with at least Triage rights and the GITHUB_REPO with the repo id you want the agent to open issues in.

Required Secrets

  • GITHUB_TOKEN
  • GITHUB_REPO

Attached Files

  • memory/PLAN.md
  • memory/INSTRUCTIONS.md