issue-agent
IssueAgent スキル - GitHub Issue分析・Label管理・自動分類。識学理論65ラベル体系による自動分類、Severity評価、影響度評価、担当者自動アサイン。Use when:- GitHub Issueを分析する時- 自動ラベル付与が必要な時- Severity/影響度を判定する時- 担当者をアサインする時- "Issue分析", "ラベル", "分類", "アサイン" がキーワードに含まれる時
When & Why to Use This Skill
IssueAgent is a specialized GitHub workflow automation tool designed to streamline issue analysis, labeling, and team assignment. By leveraging the Shikigaku theory's 65-label framework, it automatically evaluates issue severity, system impact, and technical requirements to ensure every ticket is categorized accurately and routed to the correct developer or AI agent efficiently.
Use Cases
- Automated Triage: Instantly classify incoming GitHub issues into categories such as bug, feature, refactor, or documentation to eliminate manual backlog grooming.
- Priority & Severity Assessment: Automatically determine the urgency (Sev.1-5) and impact level of new reports to ensure critical system failures are prioritized for immediate resolution.
- Smart Resource Allocation: Use CODEOWNERS integration to automatically assign the most qualified team member or specialized AI agent based on the issue's technical scope.
- Project Estimation: Generate data-driven time estimates for tasks based on issue type and complexity, improving the accuracy of sprint planning and delivery timelines.
- Standardized Labeling: Implement a consistent organizational labeling system across large-scale repositories to improve searchability and reporting metrics.
| name | issue-agent |
|---|---|
| description | | |
| allowed-tools | Read, Grep, Glob, Bash |
Issue Agent Skill
Issue分析・Label管理Agent - 識学理論65ラベル体系による自動分類。
役割
- Issue種別判定 (feature/bug/refactor/docs/test/deployment)
- Severity評価 (Sev.1-5)
- 影響度評価 (Critical/High/Medium/Low)
- 識学理論65ラベル自動付与
- 担当者自動アサイン (CODEOWNERS参照)
- 依存関係抽出 (#123形式)
- 所要時間見積もり
- Agent種別自動判定
判定ルール詳細
Issue種別判定
| キーワード | Issue種別 | Agent | 優先度 |
|---|---|---|---|
| feature/add/new/implement | feature | CodeGenAgent | Medium |
| bug/fix/error/problem/broken | bug | CodeGenAgent | High |
| refactor/cleanup/improve | refactor | CodeGenAgent | Medium |
| doc/documentation/readme | docs | CodeGenAgent | Low |
| test/spec/coverage | test | CodeGenAgent | Medium |
| deploy/release/ci/cd | deployment | DeploymentAgent | High |
Severity判定
| キーワード | Severity | 対応時間 | Label |
|---|---|---|---|
| critical/urgent/emergency/blocking | Sev.1-Critical | 即座 | Sev.1-Critical |
| high priority/asap/important/major | Sev.2-High | 24時間以内 | Sev.2-High |
| (デフォルト) | Sev.3-Medium | 1週間以内 | Sev.3-Medium |
| minor/small/trivial/typo | Sev.4-Low | 2週間以内 | Sev.4-Low |
| nice to have/enhancement | Sev.5-Trivial | 優先度低 | Sev.5-Trivial |
影響度判定
| キーワード | Impact | 説明 |
|---|---|---|
| all users/entire system/data loss | Critical | 全ユーザー影響 |
| many users/major feature | High | 主要機能影響 |
| some users/workaround exists | Medium | 一部機能影響 |
| few users/cosmetic | Low | 軽微な影響 |
識学理論65ラベル体系
ラベルカテゴリ
業務カテゴリ (Issue Type)
- feature, bug, refactor, documentation, test, deployment
深刻度 (Severity)
- Sev.1-Critical, Sev.2-High, Sev.3-Medium, Sev.4-Low, Sev.5-Trivial
影響度 (Impact)
- 影響度-Critical, 影響度-High, 影響度-Medium, 影響度-Low
責任者 (Responsibility)
- 担当-開発者, 担当-テックリード, 担当-PO, 担当-AI Agent
Agent種別 (Agent Type)
- CoordinatorAgent, CodeGenAgent, ReviewAgent, IssueAgent, PRAgent, DeploymentAgent
所要時間見積もり
| Issue種別 | 基本時間 | 調整係数 |
|---|---|---|
| feature | 120分 | large: ×2, quick: ×0.5 |
| bug | 60分 | major: ×2, minor: ×0.5 |
| refactor | 90分 | complex: ×2, simple: ×0.5 |
| docs | 30分 | - |
| test | 45分 | - |
| deployment | 30分 | - |
実行コマンド
# Issue分析実行
npm run agents:issue -- --issue 270
# 複数Issue一括分析
npm run agents:issue -- --issues 270,240,276
分析コメント出力例
## IssueAgent Analysis
**Issue Type**: bug
**Severity**: Sev.2-High
**Impact**: High
**Responsibility**: Developer
**Assigned Agent**: CodeGenAgent
**Estimated Duration**: 60 minutes
### Applied Labels
- bug
- Sev.2-High
- 影響度-High
- 担当-開発者
- CodeGenAgent
### Dependencies
- #270
エスカレーション条件
Sev.2-High → CISO
- セキュリティ関連Issue
Sev.2-High → TechLead
- アーキテクチャ設計に関わるIssue
Sev.2-High → PO
- ビジネス要件に関わるIssue
メトリクス
- 実行時間: 通常5-10秒
- Label付与精度: 95%+
- Severity判定精度: 90%+
- 依存関係抽出精度: 100%