@tinkcarlos/skillora 0.2.0

package/.claude/skills/skill-expert-skills/QUICK_NAVIGATION.md

@@ -0,0 +1,374 @@
+# 快速导航索引 (Quick Navigation Index)
+
+> 最后更新: 2025-01-17
+> 用途: 快速定位所需文档和工具
+
+---
+
+## 目录
+
+1. [按任务类型导航](#1-按任务类型导航)
+2. [按学习路径导航](#2-按学习路径导航)
+3. [核心文档速查](#3-核心文档速查)
+4. [工具命令速查](#4-工具命令速查)
+5. [常见问题快速定位](#5-常见问题快速定位)
+
+---
+
+## 1. 按任务类型导航
+
+### 1.1 创建新 Skill
+
+```
+开始 → SKILL.md (主入口)
+    ↓
+Step 0: 专家化准备 → latest-knowledge-acquisition.md
+    ↓
+Step 0: 深度研究 → deep-research-methodology.md
+    ↓
+Step 0: 领域专家化 → domain-expertise-protocol.md
+    ↓
+选择模板 → skill-templates.md
+    ↓
+初始化项目 → scripts/init_skill.py
+    ↓
+编写内容 → writing-style-guide.md
+    ↓
+验证 → quick_validate.py + universal_validate.py
+```
+
+### 1.2 优化现有 Skill
+
+```
+开始 → SKILL.md (主入口)
+    ↓
+Step 0: 获取最新知识 → latest-knowledge-acquisition.md
+    ↓
+MCP 回退方案 → mcp-fallback-strategies.md
+    ↓
+Step 0: 领域专家化 → domain-expertise-protocol.md
+    ↓
+优化内容 → skills-knowledge-base.md
+    ↓
+验证 → quick_validate.py + universal_validate.py
+    ↓
+对比官方 → scripts/diff_with_official.py
+```
+
+### 1.3 打包分发
+
+```
+完成 → SKILL.md (主入口)
+    ↓
+打包 → scripts/package_skill.py
+    ↓
+验证 → skills-knowledge-base.md
+```
+
+---
+
+## 2. 按学习路径导航
+
+### 2.1 快速上手 (30 分钟)
+
+```
+1. 阅读官方最佳实践 (10 分钟)
+   → official-best-practices.md
+
+2. 查看示例 (10 分钟)
+   → examples.md
+
+3. 选择模板开始实践 (10 分钟)
+   → skill-templates.md
+```
+
+### 2.2 深入学习 (2 小时)
+
+```
+1. 理解核心原则 (30 分钟)
+   → skills-knowledge-base.md (第 1-3 节)
+
+2. 学习领域专家化 (30 分钟)
+   → domain-expertise-protocol.md
+
+3. 掌握研究方法 (30 分钟)
+   → latest-knowledge-acquisition.md
+   → deep-research-methodology.md
+
+4. 实践创建 Skill (30 分钟)
+   → examples.md → 自己动手
+```
+
+### 2.3 成为专家 (1 周)
+
+```
+Day 1: 核心概念与原则
+  → official-best-practices.md
+  → skills-knowledge-base.md (全部)
+
+Day 2: 领域专家化流程
+  → domain-expertise-protocol.md
+  → deep-research-methodology.md
+
+Day 3: 知识获取与验证
+  → latest-knowledge-acquisition.md
+  → knowledge-validation-checklist.md
+
+Day 4: 写作与样式规范
+  → writing-style-guide.md
+  → universality-guide.md
+
+Day 5: 工具与脚本
+  → tools-guide.md
+  → 阅读 scripts/ 源码
+
+Day 6: 模式与最佳实践
+  → patterns.md
+  → examples.md (深入研究)
+
+Day 7: 领域知识库
+  → domain-knowledge/_index.md
+  → 选择相关领域深入学习
+```
+
+---
+
+## 3. 核心文档速查
+
+### 3.1 必读文档 (按优先级)
+
+| 优先级 | 文档 | 用途 | 何时阅读 |
+|--------|------|------|----------|
+| 🔴🔴🔴 | official-best-practices.md | Anthropic 官方指南 | 第一优先 |
+| 🔴🔴🔴 | latest-knowledge-acquisition.md | 知识获取协议 | 创建/优化 Skill 前 |
+| 🔴🔴 | domain-expertise-protocol.md | 领域专家化流程 | 深度研究前 |
+| 🔴 | deep-research-methodology.md | 深度研究方法 | 成为专家 |
+| 🔴 | examples.md | 3 个完整示例 | 不确定时 |
+| 🟡 | skill-templates.md | 5 个模板 | 开始创建 |
+
+### 3.2 按需阅读文档
+
+| 任务类型 | 首选文档 | 备选文档 |
+|----------|----------|----------|
+| 写 description | writing-style-guide.md (第 1 节) | official-best-practices.md |
+| 写正文 | writing-style-guide.md (第 2 节) | skills-knowledge-base.md |
+| 写 references | skills-knowledge-base.md (第 4 节) | universality-guide.md |
+| 测试模式 | patterns.md (第 3 节) | testing-patterns.md |
+| 工作流设计 | patterns.md (第 1 节) | examples.md |
+
+---
+
+## 4. 工具命令速查
+
+### 4.1 初始化与验证
+
+```bash
+# 初始化新 Skill
+python scripts/init_skill.py my-new-skill --path .claude/skills
+
+# 快速验证
+python scripts/quick_validate.py .claude/skills/my-skill
+
+# 通用性验证
+python scripts/universal_validate.py .claude/skills/my-skill
+
+# 两者都运行
+python scripts/quick_validate.py .claude/skills/my-skill && \
+python scripts/universal_validate.py .claude/skills/my-skill
+```
+
+### 4.2 分析与打包
+
+```bash
+# 触发词分析
+python scripts/analyze_trigger.py .claude/skills/my-skill
+
+# 打包 Skill
+python scripts/package_skill.py .claude/skills/my-skill ./dist
+
+# 对比官方版本
+python scripts/diff_with_official.py .claude/skills/my-skill
+
+# 升级旧版 Skill
+python scripts/upgrade_skill.py .claude/skills/my-skill
+```
+
+### 4.3 安装依赖
+
+```bash
+# 进入项目目录
+cd .claude/skills/skill-expert-skills
+
+# 安装依赖
+pip install -r scripts/requirements.txt
+
+# 或使用虚拟环境（推荐）
+python -m venv venv
+source venv/bin/activate  # Windows: venv\Scripts\activate
+pip install -r scripts/requirements.txt
+```
+
+---
+
+## 5. 常见问题快速定位
+
+### 5.1 创建 Skill 问题
+
+| 问题 | 定位文档 | 关键章节 |
+|------|----------|----------|
+| 不知道从哪里开始 | examples.md, skill-templates.md | 快速选择 |
+| description 怎么写 | writing-style-guide.md | 第 1 节 |
+| SKILL.md 写什么 | official-best-practices.md | Progressive Disclosure |
+| 需要什么知识 | domain-expertise-protocol.md | 步骤 1-2 |
+
+### 5.2 优化 Skill 问题
+
+| 问题 | 定位文档 | 关键章节 |
+|------|----------|----------|
+| 如何获取最新知识 | latest-knowledge-acquisition.md | 全部 |
+| MCP 工具不可用 | mcp-fallback-strategies.md | 全部 |
+| 如何深度研究 | deep-research-methodology.md | 全部 |
+| 领域知识在哪里 | domain-knowledge/_index.md | 列表 |
+
+### 5.3 验证问题
+
+| 问题 | 定位文档 | 关键章节 |
+|------|----------|----------|
+| quick_validate 报错 | troubleshooting.md | 第 4 节 |
+| universal_validate 报错 | troubleshooting.md | 第 4 节 |
+| SKILL.md 太长 | official-best-practices.md | Conciseness |
+| 通用性问题 | universality-guide.md | Red Flags |
+
+### 5.4 跨项目问题
+
+| 问题 | 定位文档 | 关键章节 |
+|------|----------|----------|
+| 项目路径问题 | universality-guide.md | Red Flags |
+| 项目特定内容 | universality-guide.md | 抽象化步骤 |
+| 依赖问题 | troubleshooting.md | 第 4 节 |
+
+---
+
+## 6. 快速决策树
+
+### 6.1 创建 Skill 决策树
+
+```
+┌─────────────────────────────────────────┐
+│ 创建 Skill 决策树                    │
+├─────────────────────────────────────────┤
+│                                        │
+│  第一次创建？                         │
+│    ├─ 是 → examples.md + skill-templates.md (选模板)
+│    └─ 否 → 查看现有 Skill，参考结构
+│                                        │
+│  需要专业知识？                       │
+│    ├─ 是 → domain-knowledge/_index.md (查是否已有)
+│    └─ 否 → 从 skill-templates.md 开始
+│                                        │
+│  MCP 工具可用？                      │
+│    ├─ 是 → latest-knowledge-acquisition.md
+│    └─ 否 → mcp-fallback-strategies.md
+└─────────────────────────────────────────┘
+```
+
+### 6.2 优化 Skill 决策树
+
+```
+┌─────────────────────────────────────────┐
+│ 优化 Skill 决策树                    │
+├─────────────────────────────────────────┤
+│                                        │
+│  需要最新知识？                       │
+│    ├─ 是 → latest-knowledge-acquisition.md
+│    └─ 否 → 跳过，直接优化
+│                                        │
+│  需要深度研究？                       │
+│    ├─ 是 → deep-research-methodology.md
+│    └─ 否 → 跳过，使用现有知识
+│                                        │
+│  有领域知识库？                       │
+│    ├─ 是 → domain-knowledge/xxx-expertise.md
+│    └─ 否 → 跳过，使用通用知识
+│                                        │
+│  优化后验证？                         │
+│    ├─ 是 → quick_validate.py + universal_validate.py
+│    └─ 否 → 跳过验证（不推荐）
+└─────────────────────────────────────────┘
+```
+
+---
+
+## 7. 领域知识库快速定位
+
+### 7.1 领域知识库映射
+
+| 开发任务 | 推荐知识库 | 文件 |
+|----------|------------|------|
+| Bug 修复 | Bug Fixing | domain-knowledge/bug-fixing-expertise.md |
+| 代码审查 | Code Review | domain-knowledge/code-review-expertise.md |
+| 前端开发 | Frontend | domain-knowledge/frontend-expertise.md |
+| 后端开发 | Backend | domain-knowledge/backend-expertise.md |
+| API 设计 | API Design | domain-knowledge/api-design-expertise.md |
+| 安全实现 | Security | domain-knowledge/security-expertise.md |
+| 数据库操作 | Database | domain-knowledge/database-expertise.md |
+| DevOps 部署 | DevOps | domain-knowledge/devops-expertise.md |
+
+### 7.2 领域知识库现状
+
+| 状态 | 领域 | 数量 |
+|------|------|------|
+| ✅ 已创建 | Bug 修复, 代码审查, 前端, 后端 | 4 |
+| ⚠️ 待创建 (P1) | 安全, 数据库 | 2 |
+| 📝 待创建 (P2) | DevOps, API 设计 | 2 |
+| 📝 待创建 (P3) | UI/UX 设计 | 1 |
+
+---
+
+## 8. 学习检查清单
+
+### 8.1 快速上手检查
+
+- [ ] 已阅读 official-best-practices.md
+- [ ] 已查看 examples.md
+- [ ] 已选择 skill-templates.md 中的模板
+- [ ] 已运行 init_skill.py 初始化项目
+- [ ] 已编写 frontmatter (name, description)
+- [ ] 已编写正文内容 (保持简炼)
+- [ ] 已运行 quick_validate.py 验证
+
+### 8.2 深入学习检查
+
+- [ ] 已完成一周学习路径 (见 2.3 节)
+- [ ] 已理解渐进式披露模式
+- [ ] 已掌握领域专家化流程
+- [ ] 已理解知识获取协议
+- [ ] 已掌握写作规范
+- [ ] 已创建至少一个完整 Skill
+- [ ] 已通过所有验证检查
+- [ ] 已参与代码审查或 peer review
+
+---
+
+## 9. 版本信息
+
+- 本索引版本: v1.0
+- 最后更新: 2025-01-17
+- 维护者: Claude Agent
+- 相关文档: 16 个 references 文件
+
+---
+
+## 10. 获取帮助
+
+如果本索引无法解决问题：
+
+1. **查看完整文档列表**：references/ 目录
+2. **查看知识验证**：knowledge-validation-checklist.md
+3. **查看故障排除**：troubleshooting.md
+4. **查看官方资源**：official-best-practices.md 末尾
+
+**快速反馈路径**：
+- GitHub Issues: https://github.com/anthropics/skills/issues
+- Discord: https://discord.gg/anthropic
+- 文档: https://platform.claude.com/docs/en/agents-and-tools/agent-skills/

package/.claude/skills/skill-expert-skills/SKILL.md

@@ -0,0 +1,247 @@
+---
+name: skill-expert-skills
+description: |
+  Creates, optimizes, validates, and packages Claude Agent Skills (.claude/skills/*).
+
+  ⛔ MANDATORY WORKFLOW (violation = invalid output):
+  1. READ this SKILL.md completely FIRST
+  2. Step 0: Acquire domain expertise via MCP tools (mcp__context7, mcp__exa)
+  3. Step 1-5: Follow the Decision Tree strictly
+  4. Validate with quick_validate.py + universal_validate.py
+
+  Triggers when: creating new skills, writing SKILL.md, optimizing skill descriptions,
+  improving trigger rates, validating skill structure, packaging skills for distribution.
+
+  Not for: regular programming or business logic (use domain-specific skills).
+
+license: Apache-2.0
+compatibility: Requires Python 3.8+ for validation scripts
+allowed-tools: Read Write Bash Grep Glob
+# Note: metadata below contains non-standard extension fields for local use
+metadata:
+  display_name_zh: skill专家的skills
+  language: en
+  last_updated: 2025-01-17
+  version: 1.1.0
+---
+
+# Skill Expert (Skills)
+
+Transform vague "create/optimize a Skill" requests into **triggerable, reusable,
+maintainable, verifiable** Skill packages.
+
+> **Core Principles**: 🔴 **Expertise First** | 🔴 **Latest Knowledge First** |
+> Conciseness | Universality | Progressive Disclosure | No Extra Docs
+
+---
+
+## ⛔⛔⛔ STOP - MANDATORY PRE-FLIGHT CHECK
+
+**Before ANY action, confirm you have:**
+
+| # | Checkpoint | Status |
+|---|------------|--------|
+| 1 | Read this ENTIRE SKILL.md? | ☐ |
+| 2 | Identified task type? (Create / Optimize / Validate) | ☐ |
+| 3 | Ready to execute Step 0 FIRST? (Acquire Knowledge) | ☐ |
+
+**⛔ If ANY checkbox is unchecked → STOP and complete it first!**
+
+**Workflow Sequence (MUST follow in order):**
+```
+Step 0 (Knowledge) → Step 1 (Structure) → Step 2 (Validate) → Step 3 (Refine) → Done
+```
+
+---
+
+## 🔴🔴🔴 Step 0: Become a Domain Expert (ABSOLUTE FIRST)
+
+**Whether creating or optimizing a Skill, the first step MUST be becoming an expert in that domain!**
+
+```
+User Request → Step 0a: Requirement Elicitation → Step 0b: LLM's Own Knowledge → Step 1: Acquire Latest Knowledge → Step 2: Knowledge Validation → Step 3: Knowledge Fusion → Step 4: Generate Docs → Step 5: Self-Check → Write SKILL.md
+```
+
+**🔴 Core Principle**: When knowledge conflicts, latest acquired knowledge takes precedence!
+
+→ **Requirement Elicitation**: `references/requirement-elicitation-protocol.md` (显性+隐性需求挖掘)
+→ **Full Protocol**: `references/latest-knowledge-acquisition.md` (tools/fusion/research flow)
+→ **Knowledge Validation**: `references/knowledge-validation-checklist.md` (最新性/准确性/完整性)
+→ **Docs Generation**: `references/docs-generation-workflow.md` (文档生成工作流)
+→ **Expert Protocol**: `references/domain-expertise-protocol.md`
+→ **Research Methodology**: `references/deep-research-methodology.md`
+
+## 🔴 SKILL.md Positioning (NON-NEGOTIABLE)
+
+**SKILL.md is a navigation index, not an encyclopedia!**
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│  SKILL.md SHOULD be:                                            │
+│  ✅ A table of contents scannable in 30 seconds                 │
+│  ✅ Decision tree: tells AI "what situation → which file"       │
+│  ✅ Command reference: one-line key commands/scripts            │
+│  ✅ Minimal necessary constraints/contracts                     │
+│                                                                 │
+│  SKILL.md should NOT be:                                        │
+│  ❌ Detailed knowledge base/tutorials                           │
+│  ❌ Complete protocol/specification explanations                │
+│  ❌ Long examples/code blocks                                   │
+│  ❌ Background knowledge/principle explanations                 │
+│                                                                 │
+│  → All detailed content MUST go to references/                  │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+**Conciseness Checklist** (verify before writing/modifying SKILL.md):
+- [ ] 🔴 **Is new content > 20 lines?** → Must go to references/
+- [ ] Does AI need this "every time"? → If not, move to references/
+- [ ] Can this be replaced with a one-line navigation? → Use `→ references/xxx.md`
+- [ ] Is body < 500 lines? → If over, must split (hard limit 800 lines)
+- [ ] 🔴 **Contains tech-stack specific content?** → Must abstract or move to references/
+
+## Decision Tree
+
+```
+【Create New Skill】
+   0. 🔴🔴🔴 **Acquire Latest Knowledge + Become Expert (ABSOLUTE FIRST!)** ← Skip = Invalid Skill
+      a. Priority 1 (MCP): Use MCP tools for latest docs:
+         - mcp__context7__query-docs (library/framework docs)
+         - mcp__exa__web_search_exa (deep search)
+         - mcp__exa__get_code_context_exa (code context)
+         - WebSearch (latest trends)
+      b. Priority 2 (Fallback): If MCP unavailable → Use native tools → mcp-fallback-strategies.md
+      c. Check if domain-knowledge/ has related knowledge base
+      d. Deep research → deep-research-methodology.md
+      e. Persist knowledge → domain-knowledge/[domain]-expertise.md
+      f. Pass expert self-check
+  1. python scripts/init_skill.py <name> --path .claude/skills
+  2. Write frontmatter + body (🔴 body < 500 lines!)
+  3. Detailed content → references/*.md (NOT SKILL.md)
+  4. Validate → quick_validate.py + universal_validate.py
+
+【Optimize Existing Skill】
+   0. 🔴🔴🔴 **Acquire Latest Knowledge + Add Expertise (ABSOLUTE FIRST!)** ← Skip = Invalid Skill
+      a. Priority 1 (MCP): Use MCP tools for latest docs:
+         - mcp__context7__query-docs (check for API updates)
+         - mcp__exa__web_search_exa (search latest best practices)
+         - WebSearch (check for Breaking Changes)
+      b. Priority 2 (Fallback): If MCP unavailable → Use native tools → mcp-fallback-strategies.md
+      c. Check if domain-knowledge/ has related knowledge base
+      d. Research online to fill knowledge gaps
+      e. Update knowledge base
+  1. 🔴 **Mandatory Pre-check**
+     - Current SKILL.md line count? (must < 500 recommended / < 800 hard limit)
+     - Is new content > 20 lines? → Must go to references/
+     - Is new content tech-stack specific? → Must abstract or go to references/
+  2. Conciseness check: Is SKILL.md bloated? → Immediately split to references/
+  3. Abstract → references/universality-guide.md
+  4. Collect trigger examples (3-5)
+  5. Refine description + compress body
+  6. Run validation
+
+  ⛔ Absolutely Forbidden (violation = Invalid Skill):
+     - Skip acquiring latest knowledge and write Skill directly
+     - Not using latest knowledge when conflicts exist
+     - Adding 30-line checklists directly to SKILL.md
+     - Adding example code or tables in SKILL.md
+     - Modifying without checking line count
+
+【Validate/Package Only】
+  → quick_validate.py / universal_validate.py / package_skill.py
+```
+
+## Command Reference
+
+```bash
+# Initialize
+python scripts/init_skill.py <skill-name> --path .claude/skills
+
+# Validate (required)
+python scripts/quick_validate.py .claude/skills/<skill-name>
+python scripts/universal_validate.py .claude/skills/<skill-name>
+
+# Package (optional)
+python scripts/package_skill.py .claude/skills/<skill-name> ./dist
+
+# Advanced maintenance
+python scripts/upgrade_skill.py .claude/skills/<skill-name>  # Upgrade old version Skill
+python scripts/diff_with_official.py .claude/skills/<skill-name>  # Compare with official version
+```
+
+## Key Constraints
+
+| Item | Constraint |
+|------|------------|
+| `name` | hyphen-case, ≤64 chars, must match directory name |
+| `description` | No `< >`, ≤1024 chars, **third person**, cover 3-5 triggers |
+| `license` | Optional, license name or reference to bundled file |
+| `compatibility` | Optional, ≤500 chars, environment requirements |
+| `allowed-tools` | Optional, **space-delimited** tool names (e.g., "Read Write Bash") |
+| 🔴 **SKILL.md body** | **< 500 lines** (recommended), details → references/ |
+| SKILL.md total | <500 lines (recommended) / <800 lines (hard limit) |
+| Universality | No project paths/placeholders, synthetic portable examples |
+
+## Output Contract
+
+**Required**: Updated `SKILL.md` + change summary (triggers/domains/validation results)
+
+**On-demand**: `references/` | `scripts/` | `assets/`
+
+## References Navigation
+
+| 🔴 `QUICK_NAVIGATION.md` | Quick navigation index | When starting work |
+| `mcp-fallback-strategies.md` | **MCP tools unavailable fallback** | **When MCP tools fail** |
+| 🔴 `integration-examples.md` | **Complete Skill integration examples** | When integrating multiple Skills |
+| 🔴 `mcp-tools-guide.md` | MCP tools detailed guide | When using MCP tools |
+| 🔴 `official-best-practices.md` | **Anthropic Official Best Practices** | **FIRST** as authoritative reference |
+| 🔴 `latest-knowledge-acquisition.md` | Latest knowledge acquisition protocol | **FIRST** when creating/optimizing any skill |
+| 🔴 `requirement-elicitation-protocol.md` | Requirement elicitation (显性+隐性需求) | **FIRST** when gathering requirements |
+| 🔴 `knowledge-validation-checklist.md` | Knowledge validation checklist | After knowledge acquisition |
+| 🔴 `docs-generation-workflow.md` | Docs generation workflow | When generating docs from knowledge |
+| 🔴 `mcp-fallback-strategies.md` | **MCP tools unavailable fallback** | **When MCP tools fail** |
+| 🔴 `domain-expertise-protocol.md` | Domain expertise protocol | **FIRST** when deep domain research needed |
+| 🔴 `deep-research-methodology.md` | Deep research methodology (PhD-level) | When becoming authoritative expert |
+| 🔴 `domain-knowledge-template.md` | Domain knowledge base template | When persisting knowledge |
+| `domain-knowledge/_index.md` | Index of existing knowledge bases | When checking available domain knowledge |
+| `skills-knowledge-base.md` | Complete knowledge base: format/spec/best practices | For comprehensive reference |
+| `universality-guide.md` | Universality guide | When optimizing skills, **required reading** |
+| `examples.md` | 3 complete practical examples | When unsure about structure |
+| `patterns.md` | Workflow + output + testing patterns | When designing workflows or evaluation |
+| `skill-templates.md` | 5 Skill templates | When choosing Skill type |
+| `writing-style-guide.md` | Writing style guide | When writing descriptions |
+| `troubleshooting.md` | Troubleshooting guide | When validation fails |
+
+**Official Resources**:
+- [AgentSkills.io](https://agentskills.io/) - Open standard specification (Dec 2025)
+- [Skills Overview](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview)
+- [Best Practices](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices)
+- [Anthropic Skills Repo](https://github.com/anthropics/skills) - Official examples (43k+ stars)
+
+## Definition of Done
+
+**⛔ MANDATORY: Complete ALL checkboxes before declaring done!**
+
+### Pre-Flight (Must complete FIRST)
+- [ ] 🔴🔴🔴 **Read this SKILL.md completely** before starting work
+- [ ] 🔴🔴🔴 **Identified task type** (Create / Optimize / Validate)
+- [ ] 🔴🔴🔴 **Executed Step 0** (Acquire Knowledge) before any other step
+
+### Knowledge Acquisition
+- [ ] 🔴 **Acquired latest knowledge** using MCP tools (preferred) OR native tools/manual verification (acceptable) → mcp-fallback-strategies.md
+- [ ] 🔴 **Knowledge source documented** in frontmatter/description if using fallback
+- [ ] 🔴 **Deep research completed** (knowledge base created/updated)
+- [ ] 🔴 **Expert self-check passed** (no critical unknowns)
+- [ ] 🔴 **Source date < 1 year** (ensure knowledge freshness)
+
+### Quality Gates
+- [ ] Domain expertise completed
+- [ ] description covers 3-5 triggers, **uses third person**
+- [ ] SKILL.md < 500 lines
+- [ ] quick_validate.py passed
+- [ ] universal_validate.py passed
+- [ ] No project-specific information
+
+### Self-Verification Question
+**Ask yourself**: "Did I follow Step 0 → Step 1 → ... sequence, or did I skip steps?"
+If you skipped steps → Go back and redo properly!