← ClaudeAtlas

mcp-builderlisted

Creates MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Activates on: "build an MCP server", "create MCP tools", "integrate API via MCP", "write a FastMCP server", "add an MCP server to an agent", or any request to wrap an external API as LLM-callable tools in Python or Node/TypeScript.
shipshitdev/skills · ★ 35 · AI & Automation · score 77
Install: claude install-skill shipshitdev/skills
# MCP Server Development Guide # Process ## High-Level Workflow Build an MCP server in four phases: ### Phase 1: Deep Research and Planning #### 1.1 Understand Agent-Centric Design Principles Before implementation, design tools around agent workflows: **Build for Workflows, Not Just API Endpoints:** - Do not mirror every API endpoint by default. Group endpoints into workflow tools when one user task requires multiple API calls. - Consolidate related operations (e.g., `schedule_event` that both checks availability and creates event) - List the top user workflows and map each tool to one workflow. **Optimize for Limited Context:** - Return only fields needed for the task by default. - Return high-signal information, not exhaustive data dumps - Provide "concise" vs "detailed" response format options - Default to human-readable identifiers over technical codes (names over IDs) - Add pagination, field selection, or truncation for large responses. **Design Actionable Error Messages:** - Error messages include the failed field or operation, the cause, and the next valid action. - Suggest specific next steps: "Try using filter='active_only' to reduce results" - Never return raw provider errors without a tool-level explanation. **Follow Natural Task Subdivisions:** - Tool names should reflect how humans think about tasks - Group related tools with consistent prefixes for discoverability - Design tools around natural workflows, not just API structure **Use Evaluation