feat: 初始化 SCALE OS 工程框架

- 添加 SCALE Engine 配置 (.scale/)
- 添加 OpenClaw Agent 配置 (.openclaw/)
- 添加知识文档 (AGENTS.md, TOOLS.md)
- 添加质量契约和工作流配置
- 添加 22 个工作流模板
- 添加验证脚本和门控脚本
- 添加 skills-registry 技能注册表

AGENTS.md

@@ -0,0 +1,26 @@
+# .
+
+## SCALE Engine Integration (OpenClaw)
+
+This project uses SCALE Engine for AI engineering governance via OpenClaw.
+
+### Commands
+- `scale create <type> <title>` — Create artifact
+- `scale transition <id> <action>` — Transition artifact state
+- `scale list --type Spec` — List artifacts
+- `scale role activate <role>` — Switch role
+- `scale doctor` — Health check
+
+### Workflow
+1. **Explore** → Role: explorer (Read/Grep only)
+2. **Plan** → Create Spec → refine → approve (guard: ambiguity ≤ 0.2)
+3. **Implement** → Role: implementer (Edit/Write/Bash unlocked)
+4. **Verify** → Must run tests before claiming done
+5. **Learn** → Defects → Lessons → Rules → Hooks
+
+### Rules
+- 🔴 Dangerous commands are physically blocked
+- 🔴 Hardcoded secrets are blocked on Edit/Write
+- 🟡 3 identical retries triggers brute-retry detection
+- 🟡 Claiming done without running tests is blocked
+- 🟢 All tool calls are tracked in .scale/events/