coze-skill-creator

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Coze Skill Creator

任务目标

Task Objectives

本 Skill 用于:从配置文件或需求描述创建完整的 Skill，包括工具配置、工作流编排和代码生成
能力包含:配置验证、文件生成、模板渲染、技能打包
触发条件:用户描述技能需求、提供技能配置 JSON、设计工作流结构

This Skill is used to: Create complete skills from configuration files or requirement descriptions, including tool configuration, workflow orchestration, and code generation
Capabilities include: Configuration validation, file generation, template rendering, skill packaging
Trigger conditions: Users describe skill requirements, provide skill configuration JSON, design workflow structures

前置准备

Prerequisites

依赖说明:scripts 脚本所需的依赖包及版本
```
jsonschema>=4.0.0
```
非标准文件/文件夹准备:无

Dependency description: Dependencies and versions required by scripts
```
jsonschema>=4.0.0
```
Non-standard files/folders preparation: None

操作步骤

Operation Steps

标准流程:
1. 需求分析
  - 智能体根据用户描述，理解技能的核心能力、触发场景和工具需求
  - 参考 references/config-reference.md 确定配置结构
2. 配置生成
  - 根据分析结果，生成技能配置 JSON 文件
  - 配置包含:元数据、工具列表、工作流节点、脚本定义
3. Schema 验证
  - 调用
```
scripts/validate_schema.py --config <config_path>
```
    验证配置格式
  - 若验证失败，根据错误信息修正配置
4. 文件生成
  - 调用
```
scripts/generate_skill.py --config <config_path> --output <output_dir>
```
    生成技能文件
  - 脚本自动创建 SKILL.md、scripts/、references/、assets/ 结构
5. 打包测试
  - 使用
```
package_skill(skill_dir_name=<skill_name>)
```
    工具打包技能
  - 验证 .skill 文件生成成功且内容完整
可选分支:
- 当 用户提供配置 JSON:直接跳到步骤 3 进行验证
- 当 仅需生成工作流设计:参考 references/workflow-guide.md，智能体生成工作流图和节点定义
- 当 仅需配置工具集成:参考 references/tool-reference.md，智能体生成工具配置

Standard workflow:
1. Requirement Analysis
  - The agent understands the core capabilities, trigger scenarios, and tool requirements of the skill based on the user's description
  - Refer to references/config-reference.md to determine the configuration structure
2. Configuration Generation
  - Generate a skill configuration JSON file based on the analysis results
  - Configuration includes: Metadata, tool list, workflow nodes, script definitions
3. Schema Validation
  - Call
```
scripts/validate_schema.py --config <config_path>
```
    to verify the configuration format
  - If validation fails, correct the configuration based on error messages
4. File Generation
  - Call
```
scripts/generate_skill.py --config <config_path> --output <output_dir>
```
    to generate skill files
  - The script automatically creates the structure of SKILL.md, scripts/, references/, assets/
5. Packaging and Testing
  - Use the
```
package_skill(skill_dir_name=<skill_name>)
```
    tool to package the skill
  - Verify that the .skill file is generated successfully and has complete content
Optional branches:
- When the user provides configuration JSON: Skip directly to step 3 for validation
- When only workflow design is needed: Refer to references/workflow-guide.md, the agent generates workflow diagrams and node definitions
- When only tool integration configuration is needed: Refer to references/tool-reference.md, the agent generates tool configurations

资源索引

Resource Index

必要脚本:
- scripts/validate_schema.py (用途与参数:验证技能配置 JSON 是否符合 Schema，参数:--config)
- scripts/generate_skill.py (用途与参数:根据配置生成完整技能文件，参数:--config, --output)
领域参考:
- references/config-reference.md (何时读取:生成或验证配置时)
- references/workflow-guide.md (何时读取:设计工作流时)
- references/tool-reference.md (何时读取:配置工具时)
输出资产:
- assets/templates/skill-template.md (直接用于生成 SKILL.md)
- assets/templates/python-script.py (直接用于生成 Python 脚本)
- assets/templates/schema.json (用于验证配置格式)

Required scripts:
- scripts/validate_schema.py (Purpose and parameters: Verify whether the skill configuration JSON conforms to the Schema, parameter: --config)
- scripts/generate_skill.py (Purpose and parameters: Generate complete skill files based on configuration, parameters: --config, --output)
Domain references:
- references/config-reference.md (When to read: When generating or validating configurations)
- references/workflow-guide.md (When to read: When designing workflows)
- references/tool-reference.md (When to read: When configuring tools)
Output assets:
- assets/templates/skill-template.md (Directly used to generate SKILL.md)
- assets/templates/python-script.py (Directly used to generate Python scripts)
- assets/templates/schema.json (Used to verify configuration formats)

注意事项

Notes

配置 JSON 必须符合 assets/templates/schema.json 定义的 Schema
生成的技能命名必须使用小写字母和连字符，禁止使用 -skill 后缀
工作流节点的连接关系必须保持有向无环图结构
充分利用智能体的自然语言理解能力，仅在需要结构化输出时调用脚本

The configuration JSON must conform to the Schema defined in assets/templates/schema.json
Generated skill names must use lowercase letters and hyphens, and the -skill suffix is prohibited
The connection relationship of workflow nodes must maintain a directed acyclic graph structure
Make full use of the agent's natural language understanding ability, and only call scripts when structured output is required

使用示例

Usage Examples

示例 1:从需求描述创建技能

Example 1: Create a skill from a requirement description

功能说明:用户描述"需要一个处理图片压缩的技能"
执行方式:智能体分析需求→生成配置→脚本验证→生成文件→打包
关键参数:技能名称 image-compressor、描述、工具列表

配置示例:

json

{
  "name": "image-compressor",
  "description": "压缩图片文件大小",
  "tools": [
    {
      "name": "compress",
      "type": "python",
      "parameters": [{"name": "quality", "type": "integer"}]
    }
  ]
}

Function description: User describes "Need a skill to handle image compression"
Execution method: Agent analyzes requirements → generates configuration → script validation → generates files → packaging
Key parameters: Skill name image-compressor, description, tool list

Configuration example:

json

{
  "name": "image-compressor",
  "description": "Compress image file size",
  "tools": [
    {
      "name": "compress",
      "type": "python",
      "parameters": [{"name": "quality", "type": "integer"}]
    }
  ]
}

示例 2:从配置文件生成技能

Example 2: Generate a skill from a configuration file

功能说明:用户已提供 skill-config.json 配置文件
执行方式:脚本验证→脚本生成→打包
关键参数:--config skill-config.json, --output ./my-skill

命令:

python scripts/generate_skill.py --config skill-config.json --output ./my-skill

Function description: The user has provided a skill-config.json configuration file
Execution method: Script validation → script generation → packaging
Key parameters: --config skill-config.json, --output ./my-skill

Command:

python scripts/generate_skill.py --config skill-config.json --output ./my-skill

示例 3:设计工作流

Example 3: Design a workflow

功能说明:用户需要创建包含分支逻辑的技能工作流
执行方式:智能体读取工作流指南→设计节点→生成配置→生成技能
关键参数:节点类型、连接关系、条件表达式

工作流示例:

[开始] → [条件判断] → [分支A: API调用] → [结束]
                  ↘ [分支B: 本地处理] → [结束]

Function description: The user needs to create a skill workflow with branch logic
Execution method: Agent reads workflow guide → designs nodes → generates configuration → generates skill
Key parameters: Node type, connection relationship, conditional expression

Workflow example:

[Start] → [Condition Judgment] → [Branch A: API Call] → [End]
                  ↘ [Branch B: Local Processing] → [End]