返回技能目录
人工审校软件开发

技能研究档案

mcp-developer

来源声明

Use when building, debugging, or extending MCP servers or clients that connect AI systems with external tools and data sources. Invoke to implement tool handlers, configure resource providers, set up stdio/HTTP/SSE transport layers, validate schemas with Zod or Pydantic, debug protocol compliance issues, or scaffold complete MCP server/client projects using TypeScript or Python SDKs.

保留作者的英文原始描述。它表达设计意图,不代表已验证的实际效果。

Use when building, debugging, or extending MCP servers or clients that connect AI systems with external tools and data sources. Invoke to implement tool handlers, configure resource providers, set up stdio/HTTP/SSE transport layers, validate schemas with Zod or Pydantic, debug protocol compliance issues, or scaffold complete MCP server/client projects using TypeScript or Python SDKs.

developmentreferences

研究摘要

人工审校

Skill Atlas editorial review · editorial-preview-1.0.0

用于以 TypeScript 或 Python 构建和调试 Model Context Protocol 服务器与客户端的 Skill,包含模式校验和 MCP Inspector 测试。

审阅范围

对所收集的 SKILL.md 及辅助文本/代码进行确定性静态分析。机器辅助候选和发现等待人工编辑复核;不评估动态行为或效果。

可观察能力

  • 使用 TypeScript 或 Python SDK 搭建 MCP 服务器项目。
  • 用 Zod 或 Pydantic 定义工具模式并注册资源。
  • 配置 stdio、SSE 或 HTTP 传输。
  • 用 MCP Inspector 测试协议合规性并针对失败迭代。

限制与范围

  • SDK API 会随时间变化,示例未对照当前 SDK 版本验证。
  • 要求添加认证和限流,但未给出实现。
  • 协议、SDK 和工具细节位于独立参考文件中。

适用读者

  • 将 AI 工具接入内部数据的开发者
  • 构建 MCP 服务器的团队
  • 排查协议问题的工程师

技能包含哪些内容

文件清单来自记录提交中的采集源码。路径与大小属于结构观察,不代表质量评价。

个文件
6
Markdown 文件
6
个脚本
0
个引用
5
大小
50.2 KB
结构复杂度
低

文件清单

路径类型字节
skills/mcp-developer/references/protocol.md引用4.9 KB
skills/mcp-developer/references/python-sdk.md引用9.6 KB
skills/mcp-developer/references/resources.md引用11.9 KB
skills/mcp-developer/references/tools.md引用10.2 KB
skills/mcp-developer/references/typescript-sdk.md引用7.8 KB
skills/mcp-developer/SKILL.md指令5.8 KB

SKILL.md 前置元数据

{
  "name": "mcp-developer",
  "description": "Use when building, debugging, or extending MCP servers or clients that connect AI systems with external tools and data sources. Invoke to implement tool handlers, configure resource providers, set up stdio/HTTP/SSE transport layers, validate schemas with Zod or Pydantic, debug protocol compliance issues, or scaffold complete MCP server/client projects using TypeScript or Python SDKs.",
  "license": "MIT",
  "metadata": {
    "author": "https://github.com/Jeffallan",
    "company": "https://synergetic.solutions",
    "version": "1.1.0",
    "domain": "api-architecture",
    "triggers": "MCP, Model Context Protocol, MCP server, MCP client, Claude integration, AI tools, context protocol, JSON-RPC",
    "role": "specialist",
    "scope": "implementation",
    "output-format": "code",
    "related-skills": "fastapi-expert, typescript-pro, security-reviewer, devops-engineer"
  }
}

对运行顺序的一种解读

此工作流根据源码材料重建表面顺序。除非明确记录受控评估,否则它属于暂定解读,而非真实运行轨迹。

  1. 01
  2. 02
    操作

    设计并实现

    定义资源 URI 和工具模式,注册处理器并配置传输。

  3. 03
    判断

    用 Inspector 测试

    模式或响应失败时,检查错误、修复定义或处理器并重新运行。

  4. 04
    输出

    交付服务器与配置

    提供实现、模式定义、配置和设计说明。

潜在风险面

静态发现用于提示安装前需要检查的位置,不能证明一项技能安全、恶意或有效。

有需要复核的项目

收集的文本中有 75 项确定性词法发现需要复核。发现可能仅位于文档中,不证明已执行或可被利用;仅含二进制的文件未扫描。

网络访问已观察到
文件写入已观察到
凭据使用未观察到
提示

文本描述或实现了文件修改

66 次命中

此行包含可能的文件写入或重定向操作。应检查目标路径和覆盖行为;静态审查没有写入任何文件。 该规则在收集文本中命中 66 处;此处列出 3 个代表位置。命中文本可能只是文档,分析器没有执行它。

编辑安全提示

运行前审查 npx 和 pip 包,以及生成的服务器所暴露的凭据和工具权限。

上下文可能消耗在哪里

发现、激活与资源层具有不同含义。只有使用指定官方方法并记录模型的运行时,才显示精确数值。

精确 token 计数不可用

在管线使用明确模型执行官方精确计数之前,包括文件清单在内的所有 token 值均为 null。

未使用要求的官方方法记录 token 测量。本档案不会用字符数、文件大小或估计值替代。

受控评估

没有受控评估

该技能仅具备以上标明的证据。不声称任何任务成功率、token 效率、延迟或质量优势。

尚未测试

在源码可见的情况下使用

安装命令
git clone https://github.com/Jeffallan/claude-skills.git cd claude-skills # 将该 Skill 目录复制到以下任一位置: # ~/.claude/skills/mcp-developer # .claude/skills/mcp-developer
  1. 打开固定版本源码:https://github.com/Jeffallan/claude-skills/blob/67c1961b180dc92cf08983344627109e983e8f51/skills/mcp-developer/SKILL.md。
  2. 审阅许可证、前置条件和所有捆绑的参考文件。
  3. 将 Skill 目录复制到个人或项目级 Claude Code skills 目录。
  4. 重启或刷新 Claude Code,使该 Skill 可被发现。
验证安装

让 Claude Code 搭建带 Zod 模式的工具,确认其提议用 MCP Inspector 测试。

前置条件

  • SDK API 会随时间变化,示例未对照当前 SDK 版本验证。
  • 要求添加认证和限流,但未给出实现。
  • 协议、SDK 和工具细节位于独立参考文件中。

示例提示词

代表性任务

Using the mcp-developer skill, build a TypeScript MCP server over stdio exposing a validated get_weather tool and a config resource, then test it with the MCP Inspector.

使用前检查

Read the mcp-developer skill at https://github.com/Jeffallan/claude-skills/blob/67c1961b180dc92cf08983344627109e983e8f51/skills/mcp-developer/SKILL.md and its reference files. Explain its workflow, any commands it tells the agent to run, and the main risks before I decide whether to install it.

附上限制条件的判断

优势

  • 为模式和响应格式错误提供反馈循环。
  • 提供两种语言的最小可运行示例。

注意事项

  • 脚手架使用 npx 和 pip,会下载并运行第三方包。
  • 示例使用占位的天气获取逻辑,而非真实集成。
研究结论

带有明确测试循环的 MCP 服务器开发实用起点。

从结构上保持可追溯

档案是对公开工件的限时阅读,而非背书。请通过来源链接检查确切的上游材料。

Jeffallanclaude-skillsskills/mcp-developer
来源类型
社区
提交
67c1961b180dc92cf08983344627109e983e8f51
采集时间
2026年11月20日
最近核查
2026年11月20日
上游许可证
MIT
仓库星标
12,464
内容依据

仅对提供的文件进行确定性词法解析、YAML frontmatter 校验、文件清单、来源描述的工作流程提取和启发式暂定分类。由机器生成;无 LLM、执行、动态测试或人工编辑批准。