Vscode Copilot Kit

A collection of MD files that helps interact with VSCode Copilot agent.

Installation

Note

💡 Available tools supported:

• VSCode Copilot
• Opencode
• Claudecode

Common Repo Setup

1. First time setup

Clone the repo:

cd ~ && git clone git@github.com:sonht1109/vscode-copilot-kit.git

Expose the kit subcommand from bin:

# add the kit bin folder to PATH
echo 'export PATH="$HOME/vscode-copilot-kit/bin:$PATH"' >> ~/.zshrc
source ~/.zshrc

Then cd into your service and run the subcommand:

cd project && vck setup

# setup for opencode
cd project && vck setup opencode

# setup for claudecode
cd project && vck setup claude

# setup all supported tools
cd project && vck setup all

# use symbolic links instead of copying
cd project && vck setup all --link

Available options:

Usage:
  vck <command> [options]

Commands:
  setup [target]     Run setup scripts for a target environment
  update [target]    Pull latest kit changes then run setup
  help               Show this help message

Targets:
  copilot            Setup VSCode Copilot files (default)
  opencode           Setup Opencode files
  claude             Setup Claudecode files
  all                Setup all supported tools

Options:
  -cp, --copy        Copy files (default)
  -ln, --link        Create symlinks instead of copying
  -h, --help         Show this help message

2. Update the kit

If you want to update the kit to get the latest changes, run update via the subcommand (in the service folder):

cd project && vck update

# or specific targets
cd project && vck update opencode
cd project && vck update claude
cd project && vck update all

That's it. You can start using the kit. Check the usage below to know how to use this kit.

Tip

💡 If you don't see the instruction file copilot-instructions.md, you SHOULD generate one by running:

/init

This is optional, but helps Copilot work even better!

MCP Server

This kit will use MCP server to interact with external data resource. Suggest installing JIRA + Github MCP server in VSCode.

For example: VSCode copilot (Click to expand)

{
  "inputs": [
    {
      "type": "promptString",
      "id": "github_token",
      "description": "GitHub Personal Access Token",
      "password": true
    },
    {
      "type": "promptString",
      "id": "atlassian_token",
      "description": "Atlassian API Token",
      "password": true
    }
  ],
  "servers": {
    "atlassian-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CONFLUENCE_URL",
        "-e",
        "CONFLUENCE_USERNAME",
        "-e",
        "CONFLUENCE_API_TOKEN",
        "-e",
        "JIRA_URL",
        "-e",
        "JIRA_USERNAME",
        "-e",
        "JIRA_API_TOKEN",
        "ghcr.io/sooperset/mcp-atlassian:latest"
      ],
      "env": {
        "CONFLUENCE_URL": "https://<org>.atlassian.net/wiki",
        "CONFLUENCE_USERNAME": "<username>@<org>.com", // replace your email
        "CONFLUENCE_API_TOKEN": "${env:ATLASSIAN_API_TOKEN}", // this one is configured below. Avoid hardcoding here
        "JIRA_URL": "https://<org>.atlassian.net",
        "JIRA_USERNAME": "<username>@<org>.com", // replace your email
        "JIRA_API_TOKEN": "${env:ATLASSIAN_API_TOKEN}" // this one is configured below. Avoid hardcoding here
      },
      "type": "stdio"
    },
    "sequentialthinking-mcp": {
      "command": "docker",
      "args": ["run", "--rm", "-i", "mcp/sequentialthinking"]
    }
  }
}

For example: Opencode (Click to expand)

{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "atlassian-mcp": {
      "enabled": true,
      "type": "local",
      "command": [
        "docker",
        "run",
        "-i",
        "--rm",
        "-e",
        "CONFLUENCE_URL",
        "-e",
        "CONFLUENCE_USERNAME",
        "-e",
        "CONFLUENCE_API_TOKEN",
        "-e",
        "JIRA_URL",
        "-e",
        "JIRA_USERNAME",
        "-e",
        "JIRA_API_TOKEN",
        "ghcr.io/sooperset/mcp-atlassian:latest"
      ],
      "environment": {
        "CONFLUENCE_URL": "https://<org>.atlassian.net/wiki",
        "CONFLUENCE_USERNAME": "<username>@<org>.com",
        "CONFLUENCE_API_TOKEN": "{env:ATLASSIAN_API_KEY}",
        "JIRA_URL": "https://<org>.atlassian.net",
        "JIRA_USERNAME": "<username>@<org>.com",
        "JIRA_API_TOKEN": "{env:ATLASSIAN_API_KEY}"
      }
    },
    "sequentialthinking": {
      "type": "local",
      "command": [
        "docker",
        "run",
        "--rm",
        "-i",
        "mcp/sequentialthinking"
      ]
    },
    "context7": {
      "type": "remote",
      "url": "https://mcp.context7.com/mcp",
      "headers": {
        "CONTEXT7_API_KEY": "{env:CONTEXT7_API_KEY}"
      },
      "enabled": true
    }
  }
}

For example: Claudecode (Click to expand)

{
  "mcpServers": {
    "atlassian-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "CONFLUENCE_URL",
        "-e",
        "CONFLUENCE_USERNAME",
        "-e",
        "CONFLUENCE_API_TOKEN",
        "-e",
        "JIRA_URL",
        "-e",
        "JIRA_USERNAME",
        "-e",
        "JIRA_API_TOKEN",
        "ghcr.io/sooperset/mcp-atlassian:latest"
      ],
      "env": {
        "CONFLUENCE_URL": "https://<org>.atlassian.net/wiki",
        "CONFLUENCE_USERNAME": "<username>@<org>.com",
        "CONFLUENCE_API_TOKEN": "${ATLASSIAN_API_KEY}",
        "JIRA_URL": "https://<org>.atlassian.net",
        "JIRA_USERNAME": "<username>@<org>.com",
        "JIRA_API_TOKEN": "${ATLASSIAN_API_KEY}"
      },
      "type": "stdio"
    },
    "sequentialthinking-mcp": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "mcp/sequentialthinking"
      ]
    },
    "context7-mcp": {
      "type": "http",
      "url": "https://mcp.context7.com/mcp",
      "headers": {
        "CONTEXT7_API_KEY": "${CONTEXT7_API_KEY}"
      }
    },
    "figma-mcp": {
      "url": "http://localhost:3845/mcp",
      "type": "http"
    }
  }
}

Make sure to setup your env var as well. For example, im using zsh:

echo 'export GH_TOKEN=<gh_token>' >> ~/.zshrc
echo 'export ATLASSIAN_API_KEY=<atlassian_api_key>' >> ~/.zshrc

source ~/.zshrc

Workflows

Step-by-step guides for common development tasks:

• 🚀 Feature Implementation Workflow - Plan, implement, and deliver features
• 🔍 Code Review Workflow - Review code changes before committing
• 📝 Pull Request Workflow - Create and manage pull requests
• 🎨 Figma Implementation Workflow - Implement Figma designs based on requirements or JIRA tickets

Docs

• Commands:
  • create-pr
  • review-code-changes
  • review-pr
  • plan
  • init
  • cook-plan
  • cook-figma
• Agents:
  • console-spec
  • code-reviewer
  • planner
  • jira-ticket-analyzer
• Skills:
  • backend-development
  • gh
  • sequential-thinking
  • docs-seeker
  • figma-implementation

Use your local MD

In case you want to use your local MD files along with the kit, you can create files with *.local.* in name or *-local/ folders.

Eg:

.github/agents/code-reviewer.local.agent.md
.claude/skills/project-local/SKILL.md
.opencode/commands/command.local.md

These files/folders will be preserved during the setup process and won't be overwritten by the kit.

How to contribute

Contributions are welcome! To contribute, please fork the repository, make your changes in a new branch, and submit a pull request. Follow these steps:

1. Fork the repository and clone it to your local machine and checkout a new branch.
2. If you desire to add new agents/prompts/skills/instructions/templates, please add them into .github/ folder first. Then run scripts to sync to other tools like opencode and claudecode:

bash ./migrate-tools.sh

3. If you desire to add other things which are specific to a tool, like .github/hooks, .claude/settings.json, .opencode/plugins, etc., please add them into the corresponding folder and make sure to update the setup scripts if needed.

4. Commit and create PR