Find agent skills by outcome

- Use headers

lists

- Provide tool annotations (readOnlyHint

destructiveHint

Creative

3. Be action-oriented: Start with verbs (get

list

1. Use snakecase: searchusers

create_project

- Examples: slack-mcp-server

github-mcp-server

The name should be general

descriptive of the service being integrated

- Examples: slack_mcp

github_mcp

- Streamable HTTP: For remote servers

multi-client scenarios

- Return has_more

next_offset

- stdio: For local integrations

command-line tools

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

slack-mcp-server)

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

slack_mcp)

- Example: slacksendmessage

githubcreateissue

Creative

- MCP Protocol: Start with sitemap at https://modelcontextprotocol.io/sitemap.xml

then fetch specific pages with .md suffix

- Verifiable: Single

clear answer that can be verified by string comparison

3. Question Generation: Create 10 complex

realistic questions

After implementing your MCP server

create comprehensive evaluations to test its effectiveness.

To create effective evaluations

follow the process outlined in the evaluation guide:

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

complex questions.

- 🐍 Python Guide - Module organization

dependencies

Prioritize comprehensive API coverage. List endpoints to implement

starting with the most common operations.

- ⚡ TypeScript Guide - Project structure

package.json

- Transport: Streamable HTTP for remote servers

using stateless JSON (simpler to scale and maintain

Review the service's API documentation to identify key endpoints

authentication requirements