Find agent skills by outcome

- Return pagination metadata: Include `has_more`

`next_offset`/`next_cursor`

- Use headers

lists

- Provide tool annotations (readOnlyHint

destructiveHint

1. Use snake_case: `search_users`

`create_project`

The name should be general

descriptive of the service being integrated

- Examples: `slack-mcp-server`

`github-mcp-server`

- Examples: `slack_mcp`

`github_mcp`

- stdio: For local integrations

command-line tools

- Streamable HTTP: For remote servers

multi-client scenarios

- Return `has_more`

`next_offset`

- Example: `slack_send_message`

`github_create_issue`

- Node/TypeScript: `{service}-mcp-server` (e.g.

`slack-mcp-server`)

- Python: `{service}_mcp` (e.g.

`slack_mcp`)

3. Question Generation: Create 10 complex

realistic questions

- Verifiable: Single

clear answer that can be verified by string comparison

To create effective evaluations

follow the process outlined in the evaluation guide:

After implementing your MCP server

create comprehensive evaluations to test its effectiveness.

Use evaluations to test whether LLMs can effectively use your MCP server to answer realistic

complex questions.

- [⚡ TypeScript Guide](./reference/node_mcp_server.md) - Project structure

package.json

Prioritize comprehensive API coverage. List endpoints to implement

starting with the most common operations.

- [🐍 Python Guide](./reference/python_mcp_server.md) - Module organization

dependencies

Review the service's API documentation to identify key endpoints

authentication requirements

- Language: TypeScript (high-quality SDK support and good compatibility in many execution environments e.g. MCPB. Plus AI models are good at generating TypeScript code

benefiting from its broad usage

- Transport: Streamable HTTP for remote servers

using stateless JSON (simpler to scale and maintain