Beginner LevelAI, LLMs, Embeddings & MCP
What are the 3 core primitives in Model Context Protocol (Tools, Resources, Prompts)?
❓ Interview Prompt: Explain the 3 primary capability building blocks exposed by MCP Servers to MCP Clients.
The Model Context Protocol (MCP) organizes server capabilities into 3 standardized core primitives:
1. Tools (Executable Operations):
Functions that the LLM can decide to execute (e.g. create_file, execute_sql, send_email). Model-controlled.
2. Resources (Readable Data Streams):
Data sources exposed via URIs (e.g. file:///logs/app.log, postgres://tables/users) that clients can read to supply context. Application/user-controlled.
3. Prompts (Parameterized Templates):
Reusable prompt templates exposed by the server (e.g. analyze_code_security) that users can select to initiate tasks. User-controlled.
Code Implementation Example
typescript
// MCP Primitive Capability Declaration
const serverCapabilities = {
tools: { listChanged: true }, // Executable function tools
resources: { subscribe: true }, // URI readable data sources
prompts: { listChanged: true } // User prompt templates
};💡MCP capability primitives: Tools, Resources, and Prompts.
🎯 Interview Key Takeaway
MCP defines 3 core primitives: Tools for execution, Resources for context data, and Prompts for templates.
#MCP#Tools#Resources#Prompts#Beginner
Back to all AI, LLMs, Embeddings & MCP Questions→