Nima Skill Creator

Nima Skill Creator

Overview

A professional skill creation framework that bridges the gap between user需求 discovery and technical implementation.

Core Features

1. 双语引导 (Dual-Language Guidance)

• 交互阶段: 中文引导用户需求挖掘
• 技术阶段: 英文标准文档和最佳实践

2. 渐进式披露 (Progressive Disclosure)

Level 1: Trigger metadata (name + description)
  ↓ (on trigger)
Level 2: SKILL.md body (<5k words)
  ↓ (on demand)
Level 3: Bundled resources (unlimited)

Workflow

阶段 1: 需求挖掘 (Discovery)

用户输入 → 交互式提问 → 技术规范输出

关键问题:

• Claude 应该输入什么？
• Claude 应该输出什么？
• 用户会怎么说来触发 Skill？

阶段 2: 架构蓝图 (Blueprint)

规范 → 目录结构 → 资源规划

输出:

• 目录结构 (scripts/, references/, assets/)
• 资源清单
• 工作流逻辑

阶段 3: 实现 (Implementation)

蓝图 → 代码/文档 → 验证 → 打包

技术标准 (Technical Specifications)

目录结构

skill-name/
├── SKILL.md              # Required (frontmatter + body)
├── scripts/              # Optional (executable code)
├── references/           # Optional (loaded on demand)
└── assets/               # Optional (output resources)

SKILL.md 格式

---
name: skill-name-here         # lowercase-hyphen-case, <64 chars
description: <1024 chars>     # Trigger scenarios + functionality
---

命名规范

• ✅ processing-pdfs, analyzing-spreadsheets
• ❌ helper, utils, tools

Here are 3 example Skill usage scenarios:

Scenario 1: PDF Processing Skill

Trigger: "Help me extract text and tables from PDFs" Workflow:

1. Discover: Input (PDF files), Output (extracted text/tables)
2. Blueprint: scripts/extract_pdf.py, references/pdf-libraries.md
3. Implement: Create script + reference docs

Scenario 2: Image Editor Skill

Trigger: "Edit images for my presentations" Workflow:

1. Discover: Input (image files), Output (edited images)
2. Blueprint: scripts/edit_image.py, assets/templates/
3. Implement: Create editing script + templates

Scenario 3: API Integration Skill

Trigger: "Connect Claude to my Notion workspace" Workflow:

1. Discover: Input (Notion API data), Output (structured results)
2. Blueprint: scripts/notion_api.py, references/notion-schema.md
3. Implement: Create API client + schema docs

Best Practices

简洁至上 (Conciseness)

• Claude 已经很聪明 → only add non-obvious context
• 每个 token 都要问: "Does Claude really need this?"

自由度匹配 (Freedom Matching)

渐进式披露 (Progressive Disclosure)

• SKILL.md body: <500 lines, essentials only
• Detailed content: references/ files
• No deeply nested references: one level deep only
• Long files: include table of contents

Validation Rules

Required

• ✅ SKILL.md exists with valid YAML frontmatter
• ✅ name: lowercase-hyphen-case, <64 characters
• ✅ description: includes functionality + trigger scenarios

Recommended

• 📝 scripts/ for executable code
• 📚 references/ for detailed documentation
• 🎨 assets/ for reusable templates

Avoid

• ❌ README.md, INSTALLATION_GUIDE.md, etc.
• ❌ Deeply nested references
• ❌ Duplicate information across files

For Implementation

See references/ for detailed technical guides:

• best-practices.md - Naming, patterns, quality checklist
• workflows.md - Multi-step process templates
• output-patterns.md - Output format templates
• interaction-guide.md - Interactive design patterns (Chinese)

For Agents

When this skill is triggered, the AI should:

1. Present the interactive discovery questions (Chinese) - see references/interaction-guide.md
2. Generate the technical blueprint (English)
3. Execute initialization scripts from scripts/
4. Guide implementation with referenced best practices
5. Validate with scripts/validate_skill.py
6. Package with scripts/package_skill.py