datadog-mcp: Instructions file for Codex

🤖 Agent Task Guidelines - Datadog MCP Server

This document provides specialized guidance for AI agents working on specific tasks for the Datadog MCP Server.

🚀 Quick Start: How to Use This Guide

Step 1: Identify What You Want to Do

Pick ONE task from this list:

• ✅ Write tests → Go to Testing Agent
• ✅ Add a new tool (API endpoint) → Go to Tool Builder Agent
• ✅ Create a resource (AI-friendly data view) → Go to Resource Builder Agent
• ✅ Build a prompt (AI workflow) → Go to Prompt Builder Agent
• ✅ Update documentation → Go to Documentation Agent
• ✅ Debug an issue → Go to Debug & Troubleshooting Agent

Step 2: Read Your Agent's Mission

Each agent section starts with a clear Mission statement telling you what to do.

Step 3: Copy the Template

Each agent provides a template - copy it and fill in the blanks with your specific details.

Step 4: Follow the Checklist

Go through the checklist line-by-line. Check off each item as you complete it. Don't skip any!

Step 5: Validate with Commands

Each agent provides test commands and success criteria to verify your work is correct.

---

📋 Table of Contents

1. Testing Agent
2. Tool Builder Agent
3. Resource Builder Agent
4. Prompt Builder Agent
5. Documentation Agent
6. Debug & Troubleshooting Agent

---

💡 Real-World Examples

Example 1: I Want to Add a New Tool

Scenario: You want to add a tool called get_slos() to fetch Service Level Objectives.

1. Go to Tool Builder Agent section
2. Copy the Tool Template (the @mcp.tool Python code)
3. Replace new_datadog_tool with get_slos
4. Replace the docstring and logic with SLO API calls
5. Go through Tool Building Checklist - check off each item
6. Run test: pytest tests/test_datadog_mcp_server.py -v
7. Update docs/API.md with your new tool
8. Done! Your tool is now available to AI agents