deco-incident-debugging

Analyze production incidents using past learnings and live data to identify root causes.

5|2|Updated Feb 26, 2026

Search Tags:#automation #debugging #learnings #triage #incident #war-room

Ready to run

npx skills add https://github.com/decocms/deco-start --skill deco-incident-debugging

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: deco-incident-debugging
Source: https://github.com/decocms/deco-start/tree/main/.cursor/skills/deco-incident-debugging
Command: npx skills add https://github.com/decocms/deco-start --skill deco-incident-debugging

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Rapidly accelerates incident response by leveraging past learnings and live signals to surface root-cause quickly.

Core Features & Use Cases

Interactive Mode: collaborate with engineers during active incidents for real-time diagnosis.
Headless Mode: autonomous investigation triggered by incident management systems when human input is minimal.
Learnings Database: connects symptoms to proven fixes and documented learnings for faster resolution.
Structured Outputs: generates concise incident reports and recommended next actions.

Quick Start

Provide the incident brief and trigger the triage workflow to begin rapid root-cause analysis.

Frequently Asked Questions about deco-incident-debugging

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate incident triage and debugging during a production outage?▼

Automate incident triage by feeding a structured incident brief, error logs, CDN metrics, and recent changes into the workflow. The system scores matches against a learnings index to rapidly surface symptoms, identify root causes, and propose fixes for production outages.

How does rapid incident triage work with past learnings and live data?▼

Rapid incident triage works by cross-referencing live signals like error logs and CDN metrics against a documented learnings database. It scores these matches to connect current symptoms with proven fixes, accelerating root-cause analysis without relying solely on manual investigation.

Can I use automated debugging for headless incident workflows without human input?▼

Yes, you can use headless mode for autonomous investigation triggered by incident management systems. It requires minimal human input, leveraging live data and the learnings index to generate concise incident reports and recommended next actions automatically.

What is the best way to collaborate during active incident war rooms?▼

The best way to collaborate during war rooms is using interactive mode. Engineers provide a structured incident input, and the system delivers real-time diagnosis by applying live data and past learnings to quickly identify root causes and recommend next actions.

What data do I need to provide for automated incident root-cause analysis?▼

You need to provide a structured incident input, access to a learnings index, error logs, CDN metrics, and recent changes. Supplying these live signals and historical data allows the system to accurately score matches and propose effective fixes.