triage

Analyze external monitoring alerts and create GitHub issues for unresolved items.

1|Updated Mar 28, 2026
One-click install
npx skills add https://github.com/froide-kk/michinushi --skill triage-froide-kk
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: triage
Source: https://github.com/froide-kk/michinushi/tree/main/triage
Command: npx skills add https://github.com/froide-kk/michinushi --skill triage-froide-kk

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Many projects receive noisy or duplicated alerts from external monitoring services and lack a reliable, automated way to convert those signals into tracked work items. This Skill centralizes detection of configured external sources, retrieves unresolved errors, performs duplicate checks, and orchestrates analyst-driven analysis before creating tracked tasks.

Core Features & Use Cases

  • Automatic source discovery: Detects external-sources declared in project.yml and pairs them with corresponding analyze-* Skills.
  • MCP & authentication handling: Validates MCP registration and authentication when required and surfaces next steps for reauthentication or setup.
  • Delegated analysis and issue creation: Retrieves monitoring data, delegates root-cause analysis to an analyst Agent, prompts for user approval, and creates GitHub Issues or appends local tasks with appropriate labels and project placement.
  • Resilient workflow: Performs duplicate checks, skips failing sources without aborting the entire run, and supports both GitHub and local task modes.

Quick Start

Use the triage skill to analyze all configured external monitoring sources, delegate analysis to the analyst agent, and create GitHub issues or local tasks for confirmed alerts.

Frequently Asked Questions about triage

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

FAQPage Schema
How do I automatically create GitHub issues from Sentry alerts?

To create GitHub issues from Sentry alerts, the triage skill analyzes external monitoring data, deduplicates unresolved alerts against existing issues using gh, and registers new tracked tasks in GitHub Projects.

What is the best way to triage Datadog alerts and convert them into actionable tasks?

The best way to triage Datadog alerts is to use a triage skill that detects configured external sources, delegates root-cause analysis to an analyst agent, and creates actionable tasks after user approval.

How do I deduplicate monitoring alerts against existing GitHub issues?

To deduplicate monitoring alerts against existing GitHub issues, the triage skill performs duplicate checks using the gh command before creating new issues, preventing redundant work items.

Does the triage skill work with local task management when GitHub is not configured?

Yes, the triage skill supports local task modes by appending unresolved monitoring alerts and analysis results to docs/tasks.md when GitHub Projects integration is not actively configured.

How do I handle MCP authentication checks for external monitoring sources?

To handle MCP authentication checks for external monitoring sources, the triage skill validates MCP registration and authentication, surfacing next steps for reauthentication or setup if validation fails.

What happens if one monitoring source fails during the alert triage process?

If one monitoring source fails during the alert triage process, the triage skill employs a resilient workflow that skips the failing source without aborting the entire run, ensuring other sources are still processed.