Prompts

Prompts encode domain expertise (code review checklist, incident response playbook) in the server, ensuring consistent agent behavior across clients.

MCP prompts are recipe cards in a shared kitchen — every chef (client) follows the same steps for 'how to review a PR.'

Scenario

A security MCP server exposes a 'threat-model' prompt that takes a system_description argument and returns a structured threat analysis template.

Solution

In Model Context Protocol, apply Prompts to this scenario: A security MCP server exposes a 'threat-model' prompt that takes a system_description argument and returns a structured threat analysis template. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to Prompts (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where Prompts happens in the code.

1# Prompts — minimal example2from openai import OpenAI3
4client = OpenAI()  # create API client5
6# Ask the model to explain this topic7response = client.chat.completions.create(  # core API call for Prompts8    model="gpt-4o-mini",9    messages=[10        {"role": "system", "content": "You explain prompts clearly."},11        {"role": "user", "content": f"What is prompts?"},12    ],13    temperature=0,14)15print(response.choices[0].message.content)  # show output for debugging

• Treating Prompts as a black box without evaluation
• Ignoring cost and latency in production
• Skipping error handling for prompts

Quick recap — the most important points from this module.