issue-analysis_skill

Overview

This skill analyzes GitHub Issues and infers appropriate labels from Miyabi’s 57-label system across 11 categories. It helps when creating, triaging, or prioritizing Issues by producing a consistent label set and a short rationale. Results include required, recommended, optional, and automatic labels plus escalation recommendations when applicable.

How this skill works

The skill reads the issue title, body, comments, and linked issues to extract action, technical, and urgency keywords. It applies deterministic rules to select one TYPE and one PRIORITY, conditionally assigns SEVERITY for bugs, checks SPECIAL and COMMUNITY signals, and determines HIERARCHY and STATE labels automatically. Output is a structured label set with rationale, estimated effort, agent recommendation, and escalation steps.

When to use it

• When creating a new Issue to auto-suggest labels
• During triage to standardize labels across the backlog
• For backlog prioritization and SLA assignment
• When security or cost signals may require escalation
• When you need an audit-friendly rationale for labels

Best practices

• Provide a clear title and concise description to improve keyword detection
• Reference related issues or PRs to enable hierarchy detection
• Flag security or data-sensitivity details explicitly (e.g., XSS, CVE) for proper escalation
• Review inferred TYPE and PRIORITY if the issue mixes multiple intents
• Use recommended agent suggestions to speed up resolution and handoffs

Example use cases

• Automatic labeling of a newly opened bug with severity and escalation steps
• Triaging a feature request into type:feature, priority:P2-Medium, and phase:planning
• Detecting security keywords to add security and priority:P0-Critical with CISO escalation
• Identifying good-first-issue candidates for community contributors
• Assigning hierarchy labels when issues reference parent or child items