菜单
Use when developing MCP servers to access comprehensive standards for scaffolding, packaging, sampling, documentation, testing, error handling, monitoring, and CI/CD workflows.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/sandraschi/documentation-mcp --skill mcp-standards命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | mcp-standards |
| description | Use when developing MCP servers to access comprehensive standards for scaffolding, packaging, sampling, documentation, testing, error handling, monitoring, and CI/CD workflows. |
The comprehensive standards collection for the MCP (Model Context Protocol) ecosystem. This skill provides access to all MCP standards documentation, from initial scaffolding to production deployment.
Use this skill for:
Don't use for:
Minimum: FastMCP 2.14.3+ - Required for sampling support and enhanced AI workflows.
All errors must be handled with detailed logging - no empty catch blocks allowed.
# Create new MCP server
mkdir my-mcp-server
cd my-mcp-server
# Initialize with proper structure
# (See MCP Scaffolding standards)
from fastmcp import FastMCP
# Minimum version requirement
app = FastMCP("server-name", version="1.0.0")
# Implement sampling for creative servers
@app.tool()
async def sample_workflow(iterations: int = 5):
# (See AI Sampling standards)
pass
try:
result = await risky_operation()
except SpecificError as e:
logger.error(f"Detailed error: {e}", extra={
"correlation_id": request_id,
"operation": "risky_operation"
})
# Recovery logic
Start with MCP Scaffolding for project structure.
Implement AI Sampling for creative workflows.
Apply Testing Standards and Error Handling.
Implement Monitoring and Docker Standards.
mcp-server/
├── src/mcp_server_name/
├── tests/
├── docs/
├── pyproject.toml
├── README.md
└── .mcp-standards.yaml
@app.tool()
async def example_tool(ctx, param: str) -> dict:
"""Tool with proper error handling and logging."""
try:
result = await process_data(param)
logger.info("Tool executed successfully", extra={
"tool": "example_tool",
"param_length": len(param)
})
return {"success": True, "result": result}
except Exception as e:
logger.error(f"Tool execution failed: {e}", extra={
"tool": "example_tool",
"error_type": type(e).__name__,
"correlation_id": ctx.get("correlation_id")
})
return {"error": str(e)}
This skill provides hierarchical access to thousands of lines of MCP standards documentation, loading only the relevant sections for your current development task.