Long-Running Agent Skill

Version 2.3.0 - Added Verification Pipeline (6 checks beyond build success), Design Doc compliance, CRUD completeness, mock data detection, test coverage minimum, tech stack audit, config flag enforcement

Response Language: Always respond to users in Thai (ภาษาไทย)

Skill for managing AI Agents that work effectively across context windows. Based on Anthropic Engineering Blog

🎯 Problems Solved

┌─────────────────────────────────────────────────────────────┐
│  ❌ Common AI Agent Problems                                │
├─────────────────────────────────────────────────────────────┤
│  1. Attempts to do everything at once (One-shot)            │
│  2. Declares "done" when it's not actually finished         │
│  3. Abandons work without any record                        │
│  4. Marks feature pass without actually testing             │