Drop Xanthe into Claude and get a tool-layer engineer who designs tool schemas the model can actually call correctly, and secures what those tools can reach.
Xanthe owns the layer between an agent and the systems it acts on: Model Context Protocol server design, tool and function schema authoring that models call correctly on the first attempt, argument validation and coercion, idempotency and retry semantics for side-effecting tools, error surfaces the model can recover from, auth and token scoping so a tool cannot exceed its mandate, tool-count and description budgeting when selection accuracy starts degrading, and pagination and result shaping so a tool response does not blow the context window.
What you get
-
→MCP server design and tool schema authoring
-
→Argument validation, idempotency and retry semantics
-
→Auth scoping and token passthrough hazards
-
→Tool-count budgeting and result shaping for context
📄 xanthe-mcp-tool-integration-engineer.skill
Under 2 min install
Works with Claude, ChatGPT & any AI chat
How to install
Download the .skill package, open Claude, paste SKILL.md into your Project Instructions or system prompt, describe your requirement, and Xanthe builds the answer. Includes a full worked example so you see exactly what you get.