What happens if a fix doesn't work in production?

The skill includes an auto-revert check: if a fix cannot be verified with observable proof (logs or metrics) within 30 minutes, it automatically reverts the causal commit.

How does the incident-response skill integrate with Sentry?

The skill identifies Sentry links in issue bodies, uses the Sentry MCP to pull full event context, and maps stack traces to specific commits using git log.

How does it ensure bugs don't return?

The workflow forces the creation of a failing test case before the fix is applied and includes a codification step to update regression suites and monitoring rules.

Does this skill help with post-incident reporting?

Yes, it includes a dedicated /postmortem command that generates a comprehensive, blameless report including root cause analysis and a timeline of events.

Incident Response Commander

Name: Incident Response Commander
Author: phrazzld

byphrazzld

•

Analytics & Monitoring

Orchestrates the full incident lifecycle by automating triage, reproduction tests, verified fixes, and blameless postmortems.

This skill transforms Claude into a proactive incident commander capable of managing the entire remediation pipeline for production outages and critical bugs. It automates high-pressure tasks including parsing Sentry alerts, creating dedicated fix branches, and enforcing a 'test-first' reproduction workflow to ensure fixes are valid. Beyond immediate remediation, the skill streamlines the recovery process by performing automated revert checks if fixes aren't verified, generating detailed blameless postmortems, and codifying lessons learned into regression tests and monitoring rules to prevent recurrence.

Key Features

01Sentry MCP integration for automated stack trace and breadcrumb analysis

02Blameless postmortem generation with 5 Whys and timeline reconstruction

03Automated 30-minute safety revert for unverified production changes

04Systemic prevention workflow to turn incidents into long-term monitoring rules

055 GitHub stars

06Test-driven remediation that requires reproduction tests before applying fixes

Use Cases

01Rapidly resolving critical production outages triggered by observability alerts

02Automating the creation of regression tests for complex, multi-system bugs

03Standardizing the post-incident review process across engineering teams

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add phrazzld/claude-config incident-response

For use in Claude.ai and ChatGPT

Download Skill