What problem does it solve?
Turning an existing REST API into something an LLM can call requires manually writing tool definitions, schemas, and auth wiring for every endpoint. This Skill automates that entire conversion: it ingests an OpenAPI 3.x or Swagger 2.0 document and produces a working, deployable MCP server where each API operation becomes a typed MCP tool.
Core Features & Use Cases
- Spec-to-tool generation: Dereferences the spec, maps operations to snake_case tool names, and converts OpenAPI schemas into zod schemas with descriptions, enums, and constraints preserved.
- Auth wiring: Handles apiKey, HTTP bearer, HTTP basic, and OAuth2 bearer schemes via environment variables, keeping secrets out of code and conversation.
- Test and deploy pipeline: Verifies tools with the mcp-use client CLI and inspector chat, then deploys to mcp-use cloud for use from ChatGPT or Claude connectors.
- Use Case: You have a swagger.json for your company's internal API. The Skill scaffolds a project, generates one tool per endpoint with proper zod validation, lets you test calls in the inspector, and deploys a public HTTPS MCP endpoint your team can install in Claude.
Quick Start
Ask the agent to turn your OpenAPI spec file into an MCP server, for example: build an MCP server from the attached openapi.yaml and deploy it.