v1.0.0 to v1.0.0

---

name: research-agent

display-name: Research Agent

description: 技术研究代理 - 搜索文档、Stack Overflow、GitHub获取技术方案，综合输出可执行建议

version: 1.0.0

category: knowledge

user-invocable: true

tags: [research, search, documentation, stackoverflow, comparison]

capabilities: [keyword-extraction, source-compilation, recommendation-synthesis]

tools:

- file_reader

- code_analyzer

- web_search

requires:

bins: [npm]

os: [win32, darwin, linux]

handler: ./handler.js

instructions: |

Use this skill when the user needs technical research: comparing libraries, finding

solutions to errors, evaluating dependencies, or searching documentation. For library

comparison (--compare), read both package.json files if installed and compare metadata

(version, license, dependencies count, last update). For error solving (--solve), parse

the error message, extract key terms, and provide known solution patterns for common

Node.js errors (ECONNREFUSED, ENOENT, MODULE_NOT_FOUND, ERR_MODULE_NOT_FOUND, etc.).

For evaluation (--evaluate), check if the library is in the project and assess its

maintenance status. For docs (--docs), search local project files for usage patterns.

Always provide actionable recommendations with specific commands or code snippets.

examples:

- input: "/research-agent --compare express koa"

output: "express: v4.18.2, MIT, 30 deps, mature ecosystem. koa: v2.14.2, MIT, 24 deps, modern async/await. Recommendation: express for stability, koa for modern patterns."

- input: "/research-agent --solve ECONNREFUSED 127.0.0.1:5432"

output: "Error: Connection refused to PostgreSQL. Solutions: 1) Check if PostgreSQL is running (pg_isready), 2) Verify port 5432 is not blocked, 3) Check pg_hba.conf for auth settings."

- input: "/research-agent --evaluate sharp"

output: "sharp v0.33.2: MIT license, native addon (libvips), 2.1M weekly downloads, active maintenance. Used in this project for image processing. No known CVEs."

- input: "/research-agent --docs WebRTC"

output: "Found 5 local references: webrtc-data-channel.js, P2PClient.kt, WebRTCClient.kt. Key patterns: ICE candidate handling, DataChannel 'chainlesschain-data' label."

input-schema:

type: object

properties:

mode:

type: string

enum: [compare, solve, evaluate, docs]

description: Research mode

args:

type: array

items: { type: string }

description: Arguments for the selected mode

output-schema:

type: object

properties:

query: { type: string }

mode: { type: string }

results: { type: object }

recommendations: { type: array }

searchQueries: { type: array }

model-hints:

preferred: [claude-opus-4-6, gpt-4o]

cost: free

author: ChainlessChain

license: MIT

homepage: https://github.com/nicekid1/ChainlessChain

repository: https://github.com/nicekid1/ChainlessChain

---

# Research Agent 技能

## 描述

技术研究代理，帮助用户搜索文档、比较技术方案、解决错误和评估依赖。综合本地项目信息和已知解决方案模式，输出可执行的技术建议。

## 使用方法

```

/research-agent [选项]

```

## 模式

### 库比较 (--compare)

```

/research-agent --compare express koa

```

比较两个库的:

- 版本和许可证

- 依赖数量

- 维护状态

- 生态系统成熟度

- 适用场景建议

### 错误求解 (--solve)

```

/research-agent --solve "ECONNREFUSED 127.0.0.1:5432"

```

解析错误信息:

- 提取关键错误码

- 匹配已知错误模式

- 提供分步解决方案

- 生成搜索查询建议

### 库评估 (--evaluate)

```

/research-agent --evaluate lodash

```

评估单个库:

- 是否已在项目中使用

- 包元数据分析

- 维护状态评估

- 安全漏洞检查

- 替代方案建议

### 文档搜索 (--docs)

```

/research-agent --docs WebRTC

```

搜索本地项目:

- 源代码中的使用模式

- 相关文件列表

- 代码示例摘取

- 关键实现模式

## 已知错误模式

| 错误码 | 原因 | 快速解决 |

| ---------------- | --------------- | ------------------ |

| ECONNREFUSED | 目标服务未运行 | 启动服务或检查端口 |

| ENOENT | 文件/目录不存在 | 检查路径拼写 |

| MODULE_NOT_FOUND | 模块未安装 | npm install |

| EADDRINUSE | 端口被占用 | 关闭占用进程 |

| EPERM | 权限不足 | 检查文件权限 |

| ETIMEOUT | 连接超时 | 检查网络或增加超时 |

## 输出示例

```

Research: express vs koa

========================

express koa

Version: 4.18.2 2.14.2

License: MIT MIT

Dependencies: 30 24

Stars: 62k 34k

Last Update: 2 months ago 3 months ago

Bundle Size: ~208 KB ~60 KB

Recommendations:

1. Choose express for REST APIs with rich middleware ecosystem

2. Choose koa for modern async/await patterns and smaller footprint

3. Both are production-ready and well-maintained

```