Skip to main content
Use tk with GitHub Copilot in VS Code agent mode to manage issues directly from your editor.

​
What is GitHub Copilot Agent Mode?

GitHub Copilot agent mode turns Copilot into an autonomous coding agent inside VS Code. It can execute terminal commands, edit files, and — with MCP servers — interact with external tools like tk. Agent mode is available to all VS Code users with a Copilot subscription.

​
MCP Integration

Install the ticket-mcp MCP server to give Copilot agent mode direct access to your issues.

​
Workspace Configuration

Create .vscode/mcp.json in your project root: 
{
  "servers": {
    "ticket-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": ["ticket-mcp"]
    }
  }
}

​
User-Level Configuration

To make tk available across all projects, open the Command Palette (Cmd+Shift+P / Ctrl+Shift+P) and run MCP: Open User Configuration, then add the same server entry.
uv installed via: curl -LsSf https://astral.sh/uv/install.sh | shPython version 3.10 or higher (installed automatically by uv)VS Code with GitHub Copilot extension installed

​
Verify Installation

  1. Open .vscode/mcp.json in VS Code — you should see inline Start / Stop controls above the server entry
  2. Click Start to launch the MCP server
  3. Open Copilot Chat and select Agent mode from the dropdown
  4. Click the Tools icon in the chat input to confirm ticket-mcp tools are listed

​
Enabling Agent Mode

Agent mode must be active for Copilot to use MCP tools:
  1. Open Copilot Chat (click the Copilot icon in the title bar)
  2. Select Agent from the mode dropdown at the top of the chat panel
  3. MCP tools are automatically available — no extra configuration needed
In agent mode, Copilot will automatically discover and invoke tk tools based on your prompts. You don’t need to manually select tools for each interaction.

​
Project Context

For better results, add a .github/copilot-instructions.md file to give Copilot context about your issue tracking workflow: 
# Issue Tracking

This project uses **tk (ticket-rs)** for dependency-aware issue tracking.
Issues are stored as Markdown files in `.tickets/`.

## Quick Commands

- `tk triage` - AI-powered recommendations and project health
- `tk ready` - Find issues with no blockers
- `tk create "Title" -t task -p 2` - Create issue
- `tk show <id>` - View issue details
- `tk close <id>` - Complete work
- `tk dep add <blocked> <blocker>` - Add dependency

## Status Values

- `open` - Ready to be worked on
- `in_progress` - Currently being worked on
- `blocked` - Waiting on dependencies
- `closed` - Completed

## Priority Levels

- P0 - Critical
- P1 - High
- P2 - Medium (default)
- P3 - Low
- P4 - Lowest

​
Workflow Tips

​
Starting a Session

Ask Copilot: “Run tk triage and show me what I should work on today”

​
Creating Issues

Ask Copilot: “Create a high-priority bug for the login timeout issue”

​
Closing Work

Ask Copilot: “Resolve tk-abc123, I just fixed it”

​
Dependency Management

Ask Copilot: “Show me the dependency tree for tk-xyz789”

​
Project Health

Ask Copilot: “Run tk insights and explain the bottlenecks”

​
Troubleshooting

Ensure .vscode/mcp.json exists with the correct format. Open the file in VS Code and click the Start button above the server entry. Check that uvx is in your PATH.
Agent mode requires a GitHub Copilot subscription. Verify Copilot is active in VS Code and update to the latest version. Agent mode is enabled by default in VS Code 1.99+.
Make sure you’re in Agent mode (not Ask or Edit mode). Click the Tools icon in the chat input and verify ticket-mcp tools are enabled (checked).
Enable caching with tk cache info to verify the analytics cache is working. Cached responses return in under 10ms.

​
Next Steps

CLI Reference

Complete tk command documentation

MCP Server

Deep dive into MCP server capabilities

Graph Analytics

Learn about PageRank and dependency analysis

Copilot Docs

Official GitHub Copilot documentation

​
Sources