mcp-builder

Original source / Raw SKILL.md

---
name: mcp-builder
description: Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
---

# MCP Server Development Guide

## Overview

Create MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. The quality of an MCP server is measured by how well it enables LLMs to accomplish real-world tasks.

---

# Process

## High-Level Workflow

Creating a high-quality MCP server involves four main phases:

### Phase 1: Deep Research and Planning

#### 1.1 Understand Modern MCP Design

**API Coverage vs. Workflow Tools:**
Balance comprehensive API endpoint coverage with specialized workflow tools. Workflow tools can be more convenient for specific tasks, while comprehensive coverage gives agents flexibility to compose operations. Performance varies by client - some clients benefit from code execution that combines basic tools, while others work better with higher-level workflows. When uncertain, prioritize comprehensive API coverage.

**Tool Naming and Discoverability:**
Clear, descriptive tool names help agents find the right tools quickly. Use consistent prefixes (e.g., `github_create_issue`, `github_list_repos`) and action-oriented naming.

**Context Management:**
Agents benefit from concise tool descriptions and the ability to filter/paginate results. Design tools that return focused, relevant data. Some clients support code execution which can help agents filter and process data efficiently.

**Actionable Error Messages:**
Error messages should guide agents toward solutions with specific suggestions and next steps.

#### 1.2 Study MCP Protocol Documentation

**Navigate the MCP specification:**

Start with the sitemap to find relevant pages: `https://modelcontextprotocol.io/sitemap.xml`

Then fetch specific pages with `.md` suffix for markdown format (e.g., `https://modelcontextprotocol.io/specification/draft.md`).

Key pages to review:
- Specification overview and architecture
- Transport mechanisms (streamable HTTP, stdio)
- Tool, resource, and prompt definitions

#### 1.3 Study Framework Documentation

**Recommended stack:**
- **Language**: TypeScript (high-quality SDK support and good compatibility in many execution environments. Plus AI models are good at generating TypeScript code, benefiting from its broad usage, static typing and good linting tools)
- **Transport**: Streamable HTTP for remote servers, using stateless JSON (simpler to scale and maintain). stdio for local servers.

**Load framework documentation:**

- **MCP Best Practices**: See [reference/mcp_best_practices.md](reference/mcp_best_practices.md)

**For TypeScript (recommended):**
- **TypeScript SDK**: Use WebFetch to load `https://raw.githubusercontent.com/modelcontextprotocol/typescript-sdk/main/README.md`
- See [reference/node_mcp_server.md](reference/node_mcp_server.md) for TypeScript patterns and examples

**For Python:**
- **Python SDK**: Use WebFetch to load `https://raw.githubusercontent.com/modelcontextprotocol/python-sdk/main/README.md`
- See [reference/python_mcp_server.md](reference/python_mcp_server.md) for Python patterns and examples

#### 1.4 Plan Your Implementation

**Understand the API:**
Review the service's API documentation to identify key endpoints, authentication requirements, and data models. Use web search and WebFetch as needed.

**Tool Selection:**
Prioritize comprehensive API coverage. List endpoints to implement, starting with the most common operations.

---

### Phase 2: Implementation

#### 2.1 Set Up Project Structure

See language-specific guides for project setup:
- [reference/node_mcp_server.md](reference/node_mcp_server.md) - Project structure, package.json, tsconfig.json
- [reference/python_mcp_server.md](reference/python_mcp_server.md) - Module organization, dependencies

#### 2.2 Implement Core Infrastructure

Create shared utilities:
- API client with authentication
- Error handling helpers
- Response formatting (JSON/Markdown)
- Pagination support

#### 2.3 Implement Tools

For each tool:

**Input Schema:**
- Use Zod (TypeScript) or Pydantic (Python)
- Include constraints and clear descriptions
- Add examples in field descriptions

**Output Schema:**
- Define `outputSchema` where possible for structured data
- Use `structuredContent` in tool responses (TypeScript SDK feature)
- Helps clients understand and process tool outputs

**Tool Description:**
- Concise summary of functionality
- Parameter descriptions
- Return type schema

**Implementation:**
- Async/await for I/O operations
- Proper error handling with actionable messages
- Support pagination where applicable
- Return both text content and structured data when using modern SDKs

**Annotations:**
- `readOnlyHint`: true/false
- `destructiveHint`: true/false
- `idempotentHint`: true/false
- `openWorldHint`: true/false

---

### Phase 3: Review and Test

#### 3.1 Code Quality

Review for:
- No duplicated code (DRY principle)
- Consistent error handling
- Full type coverage
- Clear tool descriptions

#### 3.2 Build and Test

**TypeScript:**
- Run `npm run build` to verify compilation
- Test with MCP Inspector: `npx @modelcontextprotocol/inspector`

**Python:**
- Verify syntax: `python -m py_compile your_server.py`
- Test with MCP Inspector

See language-specific guides for detailed testing approaches and quality checklists.

---

### Phase 4: Create Evaluations

After implementing your MCP server, create comprehensive evaluations to test its effectiveness.

**Load [reference/evaluation.md](reference/evaluation.md) for complete evaluation guidelines.**

#### 4.1 Understand Evaluation Purpose

Use evaluations to test whether LLMs can effectively use your MCP server to answer realistic, complex questions.

#### 4.2 Create 10 Evaluation Questions

To create effective evaluations, follow the process outlined in the evaluation guide:

1. **Tool Inspection**: List available tools and understand their capabilities
2. **Content Exploration**: Use READ-ONLY operations to explore available data
3. **Question Generation**: Create 10 complex, realistic questions
4. **Answer Verification**: Solve each question yourself to verify answers

#### 4.3 Evaluation Requirements

Ensure each question is:
- **Independent**: Not dependent on other questions
- **Read-only**: Only non-destructive operations required
- **Complex**: Requiring multiple tool calls and deep exploration
- **Realistic**: Based on real use cases humans would care about
- **Verifiable**: Single, clear answer that can be verified by string comparison
- **Stable**: Answer won't change over time

#### 4.4 Output Format

Create an XML file with this structure:

```xml
<evaluation>
  <qa_pair>
    <question>Find discussions about AI model launches with animal codenames. One model needed a specific safety designation that uses the format ASL-X. What number X was being determined for the model named after a spotted wild cat?</question>
    <answer>3</answer>
  </qa_pair>
<!-- More qa_pairs... -->
</evaluation>
```

---

# Reference Files

Load these resources as needed during development:

### Core MCP Documentation (Load First)
- **MCP Protocol**: Start with sitemap at `https://modelcontextprotocol.io/sitemap.xml`, then fetch specific pages with `.md` suffix
- [reference/mcp_best_practices.md](reference/mcp_best_practices.md) - Universal MCP guidelines

### SDK Documentation (Load During Phase 1/2)
- **Python SDK**: Fetch from `https://raw.githubusercontent.com/modelcontextprotocol/python-sdk/main/README.md`
- **TypeScript SDK**: Fetch from `https://raw.githubusercontent.com/modelcontextprotocol/typescript-sdk/main/README.md`

### Language-Specific Implementation Guides (Load During Phase 2)
- [reference/python_mcp_server.md](reference/python_mcp_server.md) - Complete Python/FastMCP guide
- [reference/node_mcp_server.md](reference/node_mcp_server.md) - Complete TypeScript guide

### Evaluation Guide (Load During Phase 4)
- [reference/evaluation.md](reference/evaluation.md) - Complete evaluation creation guide


---

## Referenced Files

> The following files are referenced in this skill and included for context.

### reference/mcp_best_practices.md

```markdown
# MCP Server Best Practices

## Quick Reference

### Server Naming
- **Python**: `{service}_mcp` (e.g., `slack_mcp`)
- **Node/TypeScript**: `{service}-mcp-server` (e.g., `slack-mcp-server`)

### Tool Naming
- Use snake_case with service prefix
- Format: `{service}_{action}_{resource}`
- Example: `slack_send_message`, `github_create_issue`

### Response Formats
- Support both JSON and Markdown formats
- JSON for programmatic processing
- Markdown for human readability

### Pagination
- Always respect `limit` parameter
- Return `has_more`, `next_offset`, `total_count`
- Default to 20-50 items

### Transport
- **Streamable HTTP**: For remote servers, multi-client scenarios
- **stdio**: For local integrations, command-line tools

---

## Server Naming Conventions

**Python**: Use format `{service}_mcp` (lowercase with underscores)
- Examples: `slack_mcp`, `github_mcp`, `jira_mcp`

**Node/TypeScript**: Use format `{service}-mcp-server` (lowercase with hyphens)
- Examples: `slack-mcp-server`, `github-mcp-server`, `jira-mcp-server`

---

## Tool Naming and Design

### Tool Naming

1. **Use snake_case**: `search_users`, `create_project`, `get_channel_info`
2. **Include service prefix**: Anticipate that your MCP server may be used alongside others
   - Use `slack_send_message` instead of just `send_message`
   - Use `github_create_issue` instead of just `create_issue`
3. **Be action-oriented**: Start with verbs (get, list, search, create, etc.)
4. **Be specific**: Avoid generic names that could conflict with other servers

### Tool Design

- Tool descriptions must narrowly and unambiguously describe functionality
- Descriptions must precisely match actual functionality
- Provide tool annotations (readOnlyHint, destructiveHint, idempotentHint, openWorldHint)
- Keep tool operations focused and atomic

---

## Response Formats

All tools that return data should support multiple formats:

### JSON Format (`response_format="json"`)
- Machine-readable structured data
- Include all available fields and metadata
- Consistent field names and types

### Markdown Format (`response_format="markdown"`, typically default)
- Human-readable formatted text
- Use headers, lists, and formatting for clarity
- Convert timestamps to human-readable format

---

## Pagination

For tools that list resources:

- **Always respect the `limit` parameter**
- **Implement pagination**: Use `offset` or cursor-based pagination
- **Return pagination metadata**: Include `has_more`, `next_offset`/`next_cursor`, `total_count`
- **Default to reasonable limits**: 20-50 items is typical

Example pagination response:
```json
{
  "total": 150,
  "count": 20,
  "offset": 0,
  "items": [...],
  "has_more": true,
  "next_offset": 20
}
```

---

## Transport Options

### Streamable HTTP

**Best for**: Remote servers, web services, multi-client scenarios

**Use when**:
- Serving multiple clients simultaneously
- Deploying as a cloud service
- Integration with web applications

### stdio

**Best for**: Local integrations, command-line tools

**Use when**:
- Building tools for local development environments
- Integrating with desktop applications
- Single-user, single-session scenarios

**Note**: stdio servers should NOT log to stdout (use stderr for logging)

---

## Security Best Practices

### Authentication and Authorization

**OAuth 2.1**:
- Use secure OAuth 2.1 with certificates from recognized authorities
- Validate access tokens before processing requests

**API Keys**:
- Store API keys in environment variables, never in code
- Validate keys on server startup
- Provide clear error messages when authentication fails

### Input Validation

- Sanitize file paths to prevent directory traversal
- Validate URLs and external identifiers
- Check parameter sizes and ranges
- Prevent command injection in system calls
- Use schema validation (Pydantic/Zod) for all inputs

---

## Tool Annotations

Provide annotations to help clients understand tool behavior:

| Annotation | Type | Default | Description |
|-----------|------|---------|-------------|
| `readOnlyHint` | boolean | false | Tool does not modify its environment |
| `destructiveHint` | boolean | true | Tool may perform destructive updates |
| `idempotentHint` | boolean | false | Repeated calls with same args have no additional effect |
| `openWorldHint` | boolean | true | Tool interacts with external entities |

---

## Error Handling

- Use standard JSON-RPC error codes
- Report tool errors within result objects (not protocol-level errors)
- Provide helpful, specific error messages with suggested next steps
- Don't expose internal implementation details
- Clean up resources properly on errors

Example error handling:
```typescript
try {
  const result = performOperation();
  return { content: [{ type: "text", text: result }] };
} catch (error) {
  return {
    isError: true,
    content: [{
      type: "text",
      text: `Error: ${error.message}. Try using filter='active_only' to reduce results.`
    }]
  };
}
```

```

### reference/node_mcp_server.md

```markdown
# Node/TypeScript MCP Server Implementation Guide

## Quick Reference

### Key Imports
```typescript
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { StreamableHTTPServerTransport } from "@modelcontextprotocol/sdk/server/streamableHttp.js";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
import express from "express";
import { z } from "zod";
```

### Server Initialization
```typescript
const server = new McpServer({
  name: "service-mcp-server",
  version: "1.0.0"
});
```

### Tool Registration Pattern
```typescript
server.registerTool(
  "tool_name",
  {
    title: "Tool Display Name",
    description: "What the tool does",
    inputSchema: { param: z.string() },
    outputSchema: { result: z.string() }
  },
  async ({ param }) => {
    const output = { result: `Processed: ${param}` };
    return {
      content: [{ type: "text", text: JSON.stringify(output) }],
      structuredContent: output
    };
  }
);
```

---

## Server Naming Convention

Node/TypeScript MCP servers must follow this naming pattern:
- **Format**: `{service}-mcp-server` (lowercase with hyphens)
- **Examples**: `github-mcp-server`, `jira-mcp-server`, `stripe-mcp-server`

## Project Structure

```
{service}-mcp-server/
├── package.json
├── tsconfig.json
├── README.md
├── src/
│   ├── index.ts          # Main entry point
│   ├── types.ts          # TypeScript type definitions
│   ├── tools/            # Tool implementations
│   ├── services/         # API clients and utilities
│   ├── schemas/          # Zod validation schemas
│   └── constants.ts      # Shared constants
└── dist/                 # Built JavaScript files
```

## Tool Implementation

### Tool Naming

Use snake_case for tool names with service prefix:
- `slack_send_message` not `send_message`
- `github_create_issue` not `create_issue`

### Complete Tool Example

```typescript
import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { z } from "zod";

const server = new McpServer({
  name: "example-mcp",
  version: "1.0.0"
});

const UserSearchInputSchema = z.object({
  query: z.string()
    .min(2, "Query must be at least 2 characters")
    .max(200, "Query must not exceed 200 characters")
    .describe("Search string to match against names/emails"),
  limit: z.number()
    .int()
    .min(1)
    .max(100)
    .default(20)
    .describe("Maximum results to return"),
  offset: z.number()
    .int()
    .min(0)
    .default(0)
    .describe("Number of results to skip for pagination")
}).strict();

type UserSearchInput = z.infer<typeof UserSearchInputSchema>;

server.registerTool(
  "example_search_users",
  {
    title: "Search Example Users",
    description: `Search for users in the Example system by name, email, or team.

Args:
  - query (string): Search string to match against names/emails
  - limit (number): Maximum results to return (default: 20)
  - offset (number): Number of results to skip (default: 0)

Returns:
  JSON with total count, users array, and pagination info.`,
    inputSchema: UserSearchInputSchema,
    annotations: {
      readOnlyHint: true,
      destructiveHint: false,
      idempotentHint: true,
      openWorldHint: true
    }
  },
  async (params: UserSearchInput) => {
    try {
      const data = await makeApiRequest("users/search", "GET", undefined, {
        q: params.query,
        limit: params.limit,
        offset: params.offset
      });

      const output = {
        total: data.total,
        count: data.users.length,
        offset: params.offset,
        users: data.users,
        has_more: data.total > params.offset + data.users.length
      };

      return {
        content: [{ type: "text", text: JSON.stringify(output, null, 2) }],
        structuredContent: output
      };
    } catch (error) {
      return {
        content: [{
          type: "text",
          text: `Error: ${error.message}`
        }]
      };
    }
  }
);
```

## Zod Schemas

```typescript
import { z } from "zod";

const CreateUserSchema = z.object({
  name: z.string()
    .min(1, "Name is required")
    .max(100, "Name must not exceed 100 characters"),
  email: z.string()
    .email("Invalid email format"),
  age: z.number()
    .int("Age must be a whole number")
    .min(0)
    .max(150)
}).strict();

enum ResponseFormat {
  MARKDOWN = "markdown",
  JSON = "json"
}
```

## Quality Checklist

- [ ] All tools have clear descriptions
- [ ] Input validation using Zod schemas
- [ ] Consistent error handling
- [ ] Pagination support for list operations
- [ ] Tool annotations provided
- [ ] No duplicate code
- [ ] Full TypeScript type coverage
- [ ] Builds successfully with `npm run build`

```

### reference/python_mcp_server.md

```markdown
# Python MCP Server Implementation Guide

## Quick Reference

### Key Imports
```python
from mcp.server.fastmcp import FastMCP
from pydantic import BaseModel, Field, field_validator, ConfigDict
from typing import Optional, List, Dict, Any
from enum import Enum
import httpx
```

### Server Initialization
```python
mcp = FastMCP("service_mcp")
```

### Tool Registration Pattern
```python
@mcp.tool(name="tool_name", annotations={...})
async def tool_function(params: InputModel) -> str:
    # Implementation
    pass
```

---

## Server Naming Convention

Python MCP servers must follow this naming pattern:
- **Format**: `{service}_mcp` (lowercase with underscores)
- **Examples**: `github_mcp`, `jira_mcp`, `stripe_mcp`

## Tool Implementation

### Tool Naming

Use snake_case with service prefix:
- `slack_send_message` not `send_message`
- `github_create_issue` not `create_issue`

### Complete Tool Example

```python
from pydantic import BaseModel, Field, ConfigDict
from mcp.server.fastmcp import FastMCP
from typing import Optional, List
from enum import Enum

mcp = FastMCP("example_mcp")

class ResponseFormat(str, Enum):
    MARKDOWN = "markdown"
    JSON = "json"

class UserSearchInput(BaseModel):
    model_config = ConfigDict(
        str_strip_whitespace=True,
        validate_assignment=True,
        extra='forbid'
    )

    query: str = Field(
        ...,
        description="Search string to match against names/emails",
        min_length=2,
        max_length=200
    )
    limit: Optional[int] = Field(
        default=20,
        description="Maximum results to return",
        ge=1,
        le=100
    )
    offset: Optional[int] = Field(
        default=0,
        description="Number of results to skip for pagination",
        ge=0
    )
    response_format: ResponseFormat = Field(
        default=ResponseFormat.MARKDOWN,
        description="Output format: 'markdown' or 'json'"
    )

@mcp.tool(
    name="example_search_users",
    annotations={
        "title": "Search Example Users",
        "readOnlyHint": True,
        "destructiveHint": False,
        "idempotentHint": True,
        "openWorldHint": True
    }
)
async def search_users(params: UserSearchInput) -> str:
    '''Search for users in the Example system by name, email, or team.

    Args:
        params (UserSearchInput): Validated input parameters

    Returns:
        str: JSON-formatted response containing search results
    '''
    try:
        data = await _make_api_request(
            "users/search",
            params={"q": params.query, "limit": params.limit, "offset": params.offset}
        )

        response = {
            "total": data["total"],
            "count": len(data["users"]),
            "offset": params.offset,
            "users": data["users"],
            "has_more": data["total"] > params.offset + len(data["users"])
        }

        return json.dumps(response, indent=2)
    except Exception as e:
        return _handle_api_error(e)
```

## Pydantic v2 Key Features

```python
from pydantic import BaseModel, Field, field_validator, ConfigDict

class CreateUserInput(BaseModel):
    model_config = ConfigDict(
        str_strip_whitespace=True,
        validate_assignment=True
    )

    name: str = Field(..., description="User's full name", min_length=1, max_length=100)
    email: str = Field(..., description="User's email", pattern=r'^[\w\.-]+@[\w\.-]+\.\w+$')

    @field_validator('email')
    @classmethod
    def validate_email(cls, v: str) -> str:
        return v.lower()
```

## Error Handling

```python
def _handle_api_error(e: Exception) -> str:
    if isinstance(e, httpx.HTTPStatusError):
        if e.response.status_code == 404:
            return "Error: Resource not found."
        elif e.response.status_code == 403:
            return "Error: Permission denied."
        elif e.response.status_code == 429:
            return "Error: Rate limit exceeded."
        return f"Error: API request failed with status {e.response.status_code}"
    return f"Error: {type(e).__name__}"
```

## Shared Utilities

```python
async def _make_api_request(endpoint: str, method: str = "GET", **kwargs) -> dict:
    async with httpx.AsyncClient() as client:
        response = await client.request(
            method,
            f"{API_BASE_URL}/{endpoint}",
            timeout=30.0,
            **kwargs
        )
        response.raise_for_status()
        return response.json()
```

## Quality Checklist

- [ ] All tools have comprehensive docstrings
- [ ] Input validation using Pydantic models
- [ ] Consistent error handling
- [ ] Pagination support for list operations
- [ ] Tool annotations provided
- [ ] Async/await for all I/O operations
- [ ] Full type hints
- [ ] Syntax verified with `python -m py_compile`

```

### reference/evaluation.md

```markdown
# MCP Server Evaluation Guide

## Overview

Create comprehensive evaluations to test whether LLMs can effectively use your MCP server to answer realistic, complex questions.

---

## Quick Reference

### Evaluation Requirements
- Create 10 human-readable questions
- Questions must be READ-ONLY, INDEPENDENT, NON-DESTRUCTIVE
- Each question requires multiple tool calls (potentially dozens)
- Answers must be single, verifiable values
- Answers must be STABLE (won't change over time)

### Output Format
```xml
<evaluation>
   <qa_pair>
      <question>Your question here</question>
      <answer>Single verifiable answer</answer>
   </qa_pair>
</evaluation>
```

---

## Question Guidelines

### Core Requirements

1. **Questions MUST be independent** - Each question should NOT depend on answers to other questions

2. **Questions MUST require ONLY NON-DESTRUCTIVE AND IDEMPOTENT tool use** - Should not modify state

3. **Questions must be REALISTIC, CLEAR, CONCISE, and COMPLEX** - Must require multiple tools or steps

### Complexity and Depth

4. **Questions must require deep exploration** - Multi-hop questions requiring sequential tool calls

5. **Questions may require extensive paging** - May need paging through multiple pages of results

6. **Questions must require deep understanding** - Not surface-level knowledge

7. **Questions must not be solvable with straightforward keyword search** - Use synonyms, related concepts

### Tool Testing

8. **Questions should stress-test tool return values** - Large JSON objects, various data types

9. **Questions should MOSTLY reflect real human use cases**

10. **Questions may require dozens of tool calls** - Challenges LLMs with limited context

11. **Include ambiguous questions** - Force difficult decisions while maintaining single verifiable answer

### Stability

12. **Questions must be designed so the answer DOES NOT CHANGE** - Don't ask about dynamic state

---

## Answer Guidelines

### Verification

1. **Answers must be VERIFIABLE via direct string comparison**
   - Specify output format in the QUESTION
   - Examples: "Use YYYY/MM/DD.", "Respond True or False."

### Readability

2. **Answers should prefer HUMAN-READABLE formats**
   - Names, datetimes, URLs, yes/no, true/false
   - Rather than opaque IDs (though IDs are acceptable)

### Stability

3. **Answers must be STABLE/STATIONARY**
   - Look at old, closed content
   - Create questions based on concepts unlikely to change

### Diversity

4. **Answers must be DIVERSE** - Various formats: IDs, names, timestamps, URLs, booleans

5. **Answers must NOT be complex structures** - Not lists or objects unless straightforwardly verifiable

---

## Evaluation Process

### Step 1: Documentation Inspection
- Read API documentation to understand available functionality
- Fetch additional info from web if needed

### Step 2: Tool Inspection
- List available tools in the MCP server
- Understand input/output schemas and descriptions

### Step 3: Developing Understanding
- Iterate on steps 1 & 2 until good understanding
- Think about challenging tasks to create

### Step 4: Read-Only Content Inspection
- Use the MCP tools to explore actual data
- Identify specific content for questions

### Step 5: Question Creation
- Create 10 questions based on exploration
- Verify each answer independently

---

## Example Questions

**Good Question:**
```
Find discussions about AI model launches with animal codenames. One model needed
a specific safety designation that uses the format ASL-X. What number X was being
determined for the model named after a spotted wild cat?
```
Answer: `3`

**Why it's good:**
- Requires multiple searches and exploration
- Answer is stable and verifiable
- Uses synonyms (doesn't say "leopard" directly)
- Requires understanding context

**Bad Question:**
```
How many messages are in the #general channel?
```

**Why it's bad:**
- Answer changes constantly
- Too simple (one tool call)
- Not realistic human use case

```