issue-analyze

name	issue-analyze
description	Classifies a GitHub issue's title and body text into severity, label, owner, and reason. Use when asked to analyze, classify, or label GitHub issues.

Issue Analyze

This skill provides instructions for analyzing a GitHub issue's title and body text and outputting a structured JSON classification.

Input

The input consists of a GitHub issue title and body text.

Classification Rules

1. severity: Must be strictly one of HIGH, MEDIUM, or LOW.

   • HIGH: Crashes, security issues, data loss, or breaks a core flow (auth, payments, etc.).
   • MEDIUM: Performance issues or partial breakage affecting some users.
   • LOW: Typos, cosmetic issues, or documentation.

2. label: Short category string determined using keyword matching.

   • Example keyword mappings:
     • Keywords like "login", "auth", "password" → bug/auth
     • Keywords like "slow", "timeout", "lag" → perf
     • Keywords like "readme", "docs", "guide" → docs
     • Keywords like "typo", "wording", "copy", "text content", "documentation" → docs
   • If there is no confident label match, return "uncategorized".

3. owner:

   • Return a @username guess only if a clear owner signal is present in the issue context.
   • If no owner signal is available, return "unassigned" instead of guessing.

4. reason: A clear, concise one-line explanation of the classification rationale.

Output Requirement

• The output MUST be strict JSON only.
• Do NOT wrap the output in markdown codeblocks (e.g., do not use ```json or ```).
• Do NOT include any preamble, introduction, explanation, or trailing text outside of the JSON object.

JSON Schema Structure

{
  "severity": "HIGH | MEDIUM | LOW",
  "label": "short string",
  "owner": "@username or unassigned",
  "reason": "one-line explanation"
}