MCP配置文档

Tenx MCP Setup Challenge (TRP 1)

This repository contains my work for the TRP 1 – MCP Setup Challenge for 10 Academy / Tenx.
The goal is to configure my IDE with the Tenx MCP Analysis server, design an effective rules file for my coding agent, and document how these changes affect the agent’s behaviour.

Challenge Overview

• Task 1 – Setup

  • Configure my IDE (Cursor in this repo) to connect to the Tenx MCP Analysis server.
  • Ensure the MCP connection stays active during my interactions so that my usage is logged.

• Task 2 – Research & Configure

  • Study best practices for AI coding agent rules, including:
    • Boris Cherny’s workflow and community examples.
    • General patterns for controlling agent behaviour, style, safety, and tool usage.
  • Use these insights to design and refine my rules file:
    • For Cursor: .cursor/rules/agent.mdc

• Task 3 – Documentation

  • Record what I tried, what worked, what didn’t, and the insights I gained while:
    • Connecting the Tenx MCP server.
    • Iterating on the rules file and testing the agent’s behaviour.

Repository Structure

• README.md — This file, describing the challenge and repo purpose.
• .cursor/
  • mcp.json — MCP configuration (includes Tenx MCP Analysis server).
  • rules/agent.mdc — Main rules file guiding the coding agent in Cursor.
• MCP-Setup-Report.md — Detailed documentation of:
  • What I did
  • What worked
  • What didn't work / challenges
  • Insights about how rules affect the agent
• EXPERIMENTS-LOG.md — Detailed log of rules file iterations, test results, and experimentation process showing the evolution from v1.0 to v1.3

How to Run / Use Tenx MCP in Cursor

Follow these steps to use the Tenx MCP Analysis server with this repository in Cursor:

1. Open this repo in Cursor

   • Open the folder containing this project (ten-academy-trp1-mcp-challenge) in Cursor.

2. Ensure MCP config exists

   • Check that .cursor/mcp.json exists and contains the tenxfeedbackanalytics / tenxanalysismcp server pointing to https://mcppulse.10academy.org/proxy with headers:
     • "X-Device": "windows"
     • "X-Coding-Tool": "cursor"

3. Ensure rules file exists

   • Verify .cursor/rules/agent.mdc is present. This file controls how the coding agent behaves in this repo.

4. Start the Tenx MCP server in Cursor

   • Open the MCP servers panel in Cursor (via the status bar icon or the Command Palette search for “MCP”).
   • Locate the tenxfeedbackanalytics / tenxanalysismcp server and click Start.
   • Complete the GitHub authentication flow in your browser if prompted, then return to Cursor.

5. Use the agent as normal

   • Open a file (for example README.md or agent.mdc) and ask the agent for help.
   • While you interact, the Tenx MCP Analysis server will log your interactions in the background.

Author

Abnet Melaku