trae-agent-writer

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Trae Agent Writer

Create agent definitions for tasks that need independent, isolated execution.

为需要独立、隔离执行的任务创建agent定义。

What is an Agent?

什么是Agent？

A focused, autonomous instruction set for a subagent to execute independently:

Spawned as subagents - Run with isolated context
Single-purpose - One agent, one job
Stateless - No memory between invocations
Composable - Orchestrated by parent agents/skills

Agent是供子代理独立执行的专注、自主的指令集：

以子代理形式生成 - 基于隔离上下文运行
单一用途 - 一个agent对应一项任务
无状态 - 调用之间不保留记忆
可组合 - 由父级agent/技能编排

Workflow

工作流程

1. IDENTIFY → What independent task needs an agent?
2. DEFINE   → Role, inputs, process, outputs
3. DESIGN   → Step-by-step instructions
4. CREATE   → Write agent.md following the pattern

1. 识别 → 哪些独立任务需要agent？
2. 定义 → 角色、输入、流程、输出
3. 设计 → 分步执行指令
4. 创建 → 按照模板编写agent.md

Agent Format

Agent格式

markdown

undefined

markdown

undefined

{Agent Name} Agent

{Agent名称} Agent

{One-sentence role}

{一句话角色描述}

Role

角色

{What this agent does and why}

{该agent的功能及作用}

Inputs

输入

param_name: Description
output_path: Where to save results

param_name: 描述
output_path: 结果保存路径

Process

流程

Step 1: {Action}

步骤1：{操作}

Do this
Then this

执行此操作
然后执行此操作

Step N: Write Results

步骤N：写入结果

Save to

{output_path}

保存至

{output_path}

。

Output Format

输出格式

json

{
  "field": "value"
}

json

{
  "field": "value"
}

Guidelines

指南

Be objective: Avoid bias
Cite evidence: Quote specific text

undefined

保持客观：避免偏见
引用证据：引用具体文本

undefined

Agent Locations

Agent存储位置

```
skill-name/agents/
```
- Inside orchestrating skills
```
.trae/agents/
```
- Project-level agents
```
~/.trae/agents/
```
- Global agents

```
skill-name/agents/
```
- 编排技能内部
```
.trae/agents/
```
- 项目级agent
```
~/.trae/agents/
```
- 全局agent

Good Agent Candidates

适合创建Agent的场景

Pattern	Why Agent?	Example
Grader	Needs objectivity	Output evaluator
Comparator	Blind comparison	A/B tester
Analyzer	Deep dive, isolated context	Performance analyzer
Transformer	Parallel processing	File converter
Researcher	Independent investigation	Doc researcher

Don't make agents for: Simple inline tasks, tasks needing conversation history, single-step operations.

模式	为何选择Agent？	示例
评分器	需要客观性	输出评估器
比较器	盲态比较	A/B测试器
分析器	深度分析、隔离上下文	性能分析器
转换器	并行处理	文件转换器
研究员	独立调研	文档研究员

请勿为以下场景创建Agent： 简单内联任务、需要对话历史的任务、单步操作。

Example

示例

User: "I need an agent to grade code review outputs"

📄 review-grader.md

用户："我需要一个agent来评估代码评审输出"

📄 review-grader.md

Review Grader Agent

评审评分Agent

Grade code reviews against quality expectations.

根据质量标准评估代码评审内容。

Role

角色

Assess reviews for completeness, accuracy, helpfulness. Operates blindly to prevent bias.

评估评审内容的完整性、准确性、实用性。以盲态运行以避免偏见。

Inputs

输入

review_path: Path to review file
expectations: List of expected findings
output_path: Where to save grading.json

review_path: 评审文件路径
expectations: 预期检查项列表
output_path: grading.json的保存路径

Process

流程

Step 1: Read Review

步骤1：读取评审内容

Read review file
Extract all claims

读取评审文件
提取所有评审意见

Step 2: Check Expectations

步骤2：检查预期项

For each expectation:

Search for evidence
Mark PASS/FAIL
Cite evidence

针对每个预期项：

搜索相关证据
标记通过/未通过
引用证据

Step 3: Write Results

步骤3：写入结果

Save to

{output_path}/grading.json

保存至

{output_path}/grading.json

Output Format

输出格式

json

{
  "expectations": [
    {"text": "...", "passed": true, "evidence": "..."}
  ],
  "summary": {"passed": 4, "failed": 1, "pass_rate": 0.80}
}

json

{
  "expectations": [
    {"text": "...", "passed": true, "evidence": "..."}
  ],
  "summary": {"passed": 4, "failed": 1, "pass_rate": 0.80}
}

Guidelines

指南

Be objective: Don't favor verbose or brief
Cite evidence: Quote specific text

undefined

保持客观：不偏向冗长或简洁的评审
引用证据：引用具体文本

undefined

References

参考资料

Agent Patterns - Common archetypes
Agent Template - Starter template

Agent模式 - 常见架构类型
Agent模板 - 起始模板