When to Use This Skill
Use this skill when the user has any of the following needs:
- Needs to "save prompt" or "record prompt"
- Wants to "manage prompt" or "organize prompt"
- Wants to build a "prompt library" or "prompt collection"
- Needs to "categorize" or "tag" prompts
- Requires prompts to be "automatically saved" or "batch managed"
Core Objective
Automatically identify the prompt type, save it by category into a standardized directory structure, and provide unified index management so users can quickly find and reuse high-quality prompts.
Execution Flow
Step 1: Prompt Type Identification
Objective: Accurately determine the prompt's purpose and category
Identification Criteria：
-
Technical
- Code writing, debugging, optimization
- System design, architecture discussion
- Technical Q&A, solution design
- Keywords: code, development, technical, system, program
-
Content
- Article writing, content creation
- Copywriting planning, marketing content
- Blog posts, social media content
- Keywords: writing, content, copywriting, article, creation
-
Educational
- Knowledge explanation, concept clarification
- Tutorial writing, learning guidance
- Q&A, knowledge organization
- Keywords: teaching, explanation, tutorial, learning, knowledge
-
Product
- Product design, feature planning
- User experience, interaction design
- Product analysis, competitor research
- Keywords: product, design, user experience, feature
-
General
- General tasks, cross-domain applications
- Comprehensive, multi-purpose prompts
- Prompts that cannot be clearly categorized
- Keywords: general, comprehensive, universal, multi-purpose
Step 2: Automatic File Naming
Objective: Generate standardized file names
Naming Rules：
格式：[分类前缀]_[功能描述]_[版本].md
分类前缀：
- tech_ 技术类
- content_ 内容类
- edu_ 教学类
- product_ 产品类
- general_ 通用类
Naming Examples：
tech_api_debugging_v1.md
content_blog_post_outline_v1.md
edu_concept_explanation_v1.md
product_user_flow_design_v1.md
general_task_planning_v1.md
Step 3: Structured Content Saving
Objective: Save prompt content in a standard format
File Structure Template：
# [Prompt标题]
## 分类
[技术/内容/教学/产品/通用]
## 用途描述
[简要说明prompt的用途和适用场景]
## Prompt内容
[specific content of the prompt]
## 使用示例
[给出1-2个使用示例]
## 标签
[相关标签，用逗号分隔]
## 创建时间
[YYYY-MM-DD HH:MM:SS]
## 最后更新
[YYYY-MM-DD HH:MM:SS]
## 使用次数
[记录使用次数]
## 效果评价
[记录使用效果，1-5分]
Step 4: Directory Structure Creation
Objective: Establish a standard storage directory structure
Directory Structure：
prompts/
├── README.md # 总览文档
├── index.md # 索引文件
├── technical/ # 技术类prompt
│ ├── api_debugging_v1.md
│ ├── code_review_v1.md
│ └── ...
├── content/ # 内容类prompt
│ ├── blog_post_v1.md
│ ├── social_media_v1.md
│ └── ...
├── educational/ # 教学类prompt
│ ├── concept_explain_v1.md
│ ├── tutorial_v1.md
│ └── ...
├── product/ # 产品类prompt
│ ├── user_research_v1.md
│ ├── feature_design_v1.md
│ └── ...
└── general/ # 通用类prompt
├── task_planning_v1.md
├── brainstorm_v1.md
└── ...
Step 5: Index File Update
Objective: Automatically update the global index file
Index Format：
# Prompt库索引
## 分类统计
- 技术类：X个
- 内容类：X个
- 教学类：X个
- 产品类：X个
- 通用类：X个
- 总计：X个
## 最新添加
1. [文件名] - [分类] - [创建时间] - [用途]
2. ...
## 按分类浏览
### 技术类
- [api_debugging_v1.md](technical/api_debugging_v1.md) - API调试助手
- [code_review_v1.md](technical/code_review_v1.md) - 代码审查助手
### 内容类
- [blog_post_v1.md](content/blog_post_v1.md) - 博客文章生成
- ...
## 按标签浏览
[生成热门标签云]
## 搜索提示
使用关键词搜索时，建议搜索：
- 功能关键词（如：调试、写作、设计）
- 应用场景（如：面试、营销、教学）
- 技术栈（如：Python、React、数据库）
Special Handling Mechanisms
1. Duplicate Detection
Detection Logic：
- Calculate the similarity of prompt content
- Similarity > 80% is considered a duplicate
- Prompt the user whether to overwrite or update
Handling Method：
检测到相似prompt：
- 已存在：tech_api_debugging_v1.md
- 相似度：85%
- 操作选择：
1. 覆盖现有文件
2. 创建新版本（v2）
3. 保存为不同名称
2. Version Management
Version Rules：
- Similar functionality but different content: create a new version
- Identical content: prompt as duplicate
- Similar functionality: related recommendations
3. Quality Assessment
Assessment Dimensions：
- Completeness: Whether there are clear goals and instructions
- Clarity: Whether the expression is clear and unambiguous
- Practicality: Whether it solves practical problems
- Innovativeness: Whether there are unique ideas
Rating Standards：
- ⭐⭐⭐⭐⭐ Excellent: worth saving and sharing
- ⭐⭐⭐⭐ Good: has practical value
- ⭐⭐⭐ Average: usable but can be optimized
- ⭐⭐ Poor: needs substantial improvement
- ⭐ Not recommended: not advised for use
Batch Operation Features
1. Batch Import
# 支持批量导入现有prompt
glm prompts import /path/to/prompts/
2. Batch Categorization
# 自动分类未分类的prompt
glm prompts classify --all
3. Batch Export
# 导出特定分类的prompt
glm prompts export --category technical
Usage Recommendations
1. Naming Conventions
- Use descriptive file names
- Avoid special characters and spaces
- Clearly indicate version numbers
2. Content Management
- Organize and update regularly
- Delete prompts no longer needed
- Optimize low-quality prompts
3. Sharing Mechanism
- High-quality prompts can be shared with the team
- Establish a team shared library
- Synchronize updates regularly
Quality Standards
- Classification accuracy >= 95%
- File naming compliance = 100%
- Index update timeliness = 100%
- Duplicate detection accuracy >= 90%
Last updated: 2024