MCP Prompts - Claude Code
Execute MCP prompts as commands via the slash menu.
MCP prompts are reusable prompt templates an MCP server publishes. Claude Code surfaces them as slash commands so you can invoke them directly.
What it does
A server declares a prompt with a name, description, and argument schema. Claude Code registers it as /server:prompt and the user can invoke it like any other command. The server gets to shape how the prompt expands - it can fetch data, interpolate arguments, and return a structured prompt body back to Claude.
When to use it
- Shared team workflows encoded at the server level.
- Promoting common patterns as first-class commands.
- Integrating with specialized services that benefit from their own prompt shape.
- Centralizing prompt logic outside the client config.
Gotchas
- Prompt collisions across servers need namespacing. Use the server prefix.
- Servers can change prompt behavior server-side - expectations drift.
- Treat MCP prompts like skills in terms of safety review.
Official docs: https://code.claude.com/docs/en/mcp.md#use-mcp-prompts-as-commands
Technical content at the intersection of AI and development. Building with AI agents, Claude Code, and modern dev tools - then showing you exactly how it works.
Get the weekly deep dive
Tutorials on Claude Code, AI agents, and dev tools - delivered free every week.
Was this helpful?
