agent-creator

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Agent Profile Creator

Agent配置文件生成器

Create agent profiles stored as
config.json
in
~/.aider-desk/agents/{name}/
(global) or
{project}/.aider-desk/agents/{name}/
(project-level).
READ REFERENCES before proposing to ensure accuracy:
  • Tool groups & schema:
    references/agent-profile-schema.md
  • Tool approvals:
    references/tool-approval-guide.md
    +
    references/agent-profile-schema.md
  • Subagent config:
    references/subagent-guide.md
  • Examples:
    references/profile-examples.md
创建的Agent配置文件将存储为
config.json
,路径为
~/.aider-desk/agents/{name}/
(全局级)或
{project}/.aider-desk/agents/{name}/
(项目级)。
在生成配置前请阅读参考文档,以确保准确性:
  • 工具组与 schema:
    references/agent-profile-schema.md
  • 工具权限审批:
    references/tool-approval-guide.md
    +
    references/agent-profile-schema.md
  • 子Agent配置:
    references/subagent-guide.md
  • 示例:
    references/profile-examples.md

Simplified Q&A Process

简化的问答流程

Step 1: Understand Purpose

步骤1:明确用途

Ask user: "Describe the agent's purpose and what it should do."
Based on their response, internally propose all properties:
  • Name (derived from purpose)
  • Max iterations (default 250)
  • Tool groups (based on purpose)
  • Custom instructions (if applicable)
  • Subagent config (always enabled, contextMemory: "off" by default)
  • Tool approvals (based on needs)
READ all reference files before proposing.
询问用户:"请描述该Agent的用途以及它需要执行的任务。"
根据用户的回复,在内部生成所有配置属性:
  • 名称(从用途推导)
  • 最大迭代次数(默认250)
  • 工具组(基于用途选择)
  • 自定义指令(如有需要)
  • 子Agent配置(始终启用,contextMemory默认值为"off")
  • 工具权限审批(根据需求设置)
生成前请阅读所有参考文件

Step 2: Provider/Model

步骤2:提供商/模型

Ask: "Which provider/model? (format: provider/model, e.g., anthropic/claude-sonnet-4-5-20250929)"
  • Split by first slash to get provider and model
  • No validation - use as-is (correct obvious typos only)
询问:"使用哪个提供商/模型?(格式:提供商/模型,例如:anthropic/claude-sonnet-4-5-20250929)"
  • 按第一个斜杠拆分,获取提供商和模型信息
  • 不做验证:直接使用用户输入(仅修正明显的拼写错误)

Step 3: Optional Advanced Settings

步骤3:可选高级设置

ONLY ask if user specifically mentioned temperature, max tokens, rate limiting, or MCP servers:
  • "What temperature should I use?" (0.1-1.0, optional)
  • "Max tokens limit?" (optional)
  • "Min time between tool calls in ms?" (optional)
  • "Which MCP servers should be enabled?" (optional, default: none)
If not mentioned, skip this entire step.
仅当用户明确提及温度参数、最大令牌数、速率限制或MCP服务器时才询问:
  • "需要设置什么温度值?"(范围0.1-1.0,可选)
  • "最大令牌数限制是多少?"(可选)
  • "工具调用之间的最小间隔时间(毫秒)?"(可选)
  • "需要启用哪些MCP服务器?"(可选,默认:无)
如果用户未提及,跳过此步骤。

Step 4: Location

步骤4:存储位置

Ask: "Global profile (all projects) or project-specific? (default: global)"
询问:"配置文件存储为全局(所有项目可用)还是项目专属?(默认:全局)"

Step 5: Single Review & Confirm

步骤5:一次性审核与确认

Present one complete summary with all proposed properties. Ask:
"Here's your agent profile. Should I create it?"
Show:
  • Name, provider/model
  • Tool groups enabled
  • Tool approvals summary (which are "never", which are "always")
  • Subagent config
  • Location
DO NOT ask for confirmations on individual items. Only one final approval.
展示完整的配置摘要,包含所有生成的属性。询问:
"以下是您的Agent配置文件,是否确认创建?"
展示内容包括:
  • 名称、提供商/模型
  • 已启用的工具组
  • 工具权限审批摘要(哪些设置为"never",哪些为"always")
  • 子Agent配置
  • 存储位置
请勿针对单个配置项逐一确认。仅进行一次最终确认。

Step 6: Create

步骤6:创建配置文件

On user confirmation, generate the profile.
READ
references/profile-examples.md
to verify structure before creating files.
用户确认后,生成配置文件。
在创建文件前请阅读
references/profile-examples.md
,以验证配置结构。

Tool Group Reference

工具组参考

  • Power Tools: File system, shell, search (for agents that read/write files)
  • Aider Tools: AI code generation via Aider (NOT default - see note below)
  • Power Search: Advanced code search (for analysis/docs)
  • Todo Tools: Task list management (for agents that track work)
  • Memory Tools: Persistent information storage (for learning agents)
  • Skills Tools: Access to project-specific skills (for agents using custom skills)
  • Subagents: Task delegation to specialized subagents (for agents that can call others)
Aider Tools Note: Specialized for AI code generation. Do NOT enable by default.
  • If user wants coding: explain Aider (advanced refactoring) vs Power Tools (direct editing)
  • Ask preference before enabling
  • Power Tools(强力工具组):文件系统、Shell、搜索(适用于需要读写文件的Agent)
  • Aider Tools(Aider工具组):通过Aider实现AI代码生成(默认不启用 - 见下方说明)
  • Power Search(高级搜索工具组):高级代码搜索(适用于分析/文档类Agent)
  • Todo Tools(任务管理工具组):任务列表管理(适用于需要跟踪工作的Agent)
  • Memory Tools(记忆工具组):持久化信息存储(适用于具备学习能力的Agent)
  • Skills Tools(技能工具组):访问项目专属技能(适用于使用自定义技能的Agent)
  • Subagents(子Agent工具组):将任务委派给专业子Agent(适用于可调用其他Agent的Agent)
Aider Tools说明:专为AI代码生成设计。默认不启用。
  • 如果用户需要编码功能:向其说明Aider(高级重构)与Power Tools(直接编辑)的区别
  • 确认用户偏好后再启用

Tool Approval Strategy

工具权限审批策略

Keys:
{group}---{tool}
(three dashes)
Default is "ask". Only set:
  • "never": Tools completely irrelevant (e.g.,
    power---bash
    for read-only agents)
  • "always": Safe, essential tools (e.g., read operations for reviewers)
READ
references/tool-approval-guide.md
to see all available tools and their purposes.
Only include tools that exist in reference docs.
键名格式:
{group}---{tool}
(三个短横线)
默认值为"ask"。仅在以下情况修改:
  • "never":完全无关的工具(例如,只读型Agent的
    power---bash
  • "always":安全且必要的工具(例如,审核类Agent的只读操作工具)
请阅读
references/tool-approval-guide.md
,查看所有可用工具及其用途。
仅包含参考文档中存在的工具。

Subagent Configuration

子Agent配置

Every agent is a subagent (enabled: true)
READ
references/subagent-guide.md
for detailed guidance on context memory modes and configuration:
  • contextMemory
    : Default is
    "off"
    (fresh each time)
    • Use
      "full-context"
      only for specialized analysis agents (code review, security audit)
    • Use
      "last-message"
      for progressive builders (documentation, iterative tasks)
  • systemPrompt
    : Specialized for agent's purpose
  • invocationMode
    : automatic (specialized), on-demand (general)
  • color
    : Relevant color (e.g., red=security, blue=power tools)
  • description
    : Clear description for auto-invocation
所有Agent均为子Agent(启用状态:true)
请阅读
references/subagent-guide.md
,获取关于上下文记忆模式和配置的详细指导:
  • contextMemory
    默认值为
    "off"
    (每次启动均为全新上下文)
    • 仅针对专业分析类Agent(代码审核、安全审计)使用
      "full-context"
    • 针对渐进式构建类Agent(文档编写、迭代任务)使用
      "last-message"
  • systemPrompt
    :根据Agent用途定制的系统提示词
  • invocationMode
    :automatic(专业型)、on-demand(通用型)
  • color
    :相关颜色(例如:红色=安全类、蓝色=强力工具组类)
  • description
    :清晰的描述,用于自动调用

File Structure

文件结构

{profile-name}/
└── config.json
{profile-name}/
└── config.json

Validation

验证规则

  • Unique name
  • Provider/model: Use as-is (correct obvious typos only)
  • Tool keys:
    {group}---{tool}
    format
  • Values: "always", "ask", or "never"
  • Subagent enabled
  • All tool keys from references
  • 配置文件名称唯一
  • 提供商/模型:直接使用用户输入(仅修正明显拼写错误)
  • 工具键名:符合
    {group}---{tool}
    格式
  • 权限值:只能是"always"、"ask"或"never"
  • 子Agent已启用
  • 所有工具键名均来自参考文档

Resources

资源

  • references/agent-profile-schema.md
    - Complete schema
  • references/subagent-guide.md
    - Subagent configuration guide (context memory modes)
  • references/profile-examples.md
    - Examples
  • references/tool-approval-guide.md
    - Tool approval config
  • assets/templates/config.json.template
    - Template
  • assets/examples/sample-profile.json
    - Example
  • references/agent-profile-schema.md
    - 完整配置 schema
  • references/subagent-guide.md
    - 子Agent配置指南(上下文记忆模式)
  • references/profile-examples.md
    - 配置示例
  • references/tool-approval-guide.md
    - 工具权限配置指南
  • assets/templates/config.json.template
    - 配置模板
  • assets/examples/sample-profile.json
    - 示例配置文件