Connecting GitHub Copilot CLI
- Last Updated: November 20, 2026
- 3 minute read
- Semaphore
- Version 5.13
- Documentation
You can connect GitHub Copilot CLI to the Semaphore MCP Server to explore and manage Knowledge Model Management (KMM) resources using natural-language prompts.
The Semaphore MCP Server uses Streamable HTTP transport. Authentication depends on the deployment environment:
- On-premises: Use a Semaphore access token in the
X-Api-KeyHTTP header. - Progress Data Cloud: Use a JWT token in the
AuthorizationHTTP header with theBearerscheme.
This guide explains how to configure the connection, verify that the MCP Server is available, and start using Semaphore tools in GitHub Copilot CLI.
Before you begin
Make sure you have:
- Access to Semaphore 5.13.0 or later with the MCP Server available.
- GitHub Copilot CLI installed on your machine.
- The MCP Server endpoint URL for your environment.
- Network access to the Semaphore MCP Server.
- A valid token with the permissions required for your intended KMM operations: a Semaphore access token for on-premises, or a JWT token for Progress Data Cloud.
For connection requirements, see Prerequisites and Server Configuration.
For token information, see Authentication and Permissions. For instructions on generating a Progress Data Cloud JWT token, see REST API Integration.
Step 1: Find your MCP Server endpoint
For on-premises deployments, the Semaphore MCP Server endpoint follows this format:
http://<semaphore_studio_url>/ai/mcp
Replace <semaphore_studio_url> with the hostname and port of your Semaphore Studio instance.
For example, if Semaphore Studio is accessible at:
http://semaphore.example.com:5080
the MCP Server endpoint is:
http://semaphore.example.com:5080/ai/mcp
Use the protocol, hostname, and port configured for your Semaphore environment.
For Progress Data Cloud deployments, use the MCP Server endpoint URL provided for your cloud environment.
Step 2: Configure GitHub Copilot CLI
GitHub Copilot CLI may request OAuth authentication when you attempt to add the Semaphore MCP Server interactively using /mcp add.
Semaphore MCP authentication uses an HTTP token header: X-Api-Key for on-premises deployments, or Authorization with a Bearer token for Progress Data Cloud. If interactive setup requests OAuth, use the manual configuration method described below.
Locate the GitHub Copilot CLI MCP configuration file:
~/.copilot/mcp-config.jsonIf the file does not exist, create it in the
.copilotdirectory in your user home directory.For on-premises deployments, add the following configuration:
{ "mcpServers": { "MCP_Knowledge_Model_Management": { "tools": ["*"], "type": "http", "url": "http://<semaphore_studio_url>/ai/mcp", "headers": { "X-Api-Key": "<API_KEY>" }, "description": "Semaphore Knowledge Model Management MCP Server" } } }For Progress Data Cloud deployments, use your cloud MCP Server endpoint URL in the
urlfield and replace theheadersobject with:"headers": { "Authorization": "Bearer <cloud_token>" }To generate a JWT token, see REST API Integration.
Replace the example endpoint with the MCP Server endpoint for your deployment.
Replace the token placeholder with the appropriate token:
- On-premises: Replace
<API_KEY>with the access token generated in Semaphore Studio. - Progress Data Cloud: Replace
<cloud_token>with your generated JWT token.
- On-premises: Replace
Save the configuration file.
Restart GitHub Copilot CLI to load the updated MCP Server configuration.
Important: If your configuration file already contains other MCP servers, add the Semaphore server to the existing mcpServers object instead of replacing the entire file.
Protect your access token
The token provides access to Semaphore according to the permissions of the associated user.
- Do not share your token.
- Do not commit configuration files containing tokens to source control.
- Store the configuration file securely.
- Generate a new token when the current token expires.
For more information, see Authentication and Permissions.
Step 3: Verify the connection
After restarting GitHub Copilot CLI:
Start a GitHub Copilot CLI session.
Enter:
/mcpLocate
MCP_Knowledge_Model_Managementin the list of configured servers.Verify that the server is connected.
Confirm that its MCP tools are available.
If the server is not connected, see Limitations and Troubleshooting.
Step 4: Work with knowledge models
Use natural-language prompts to retrieve knowledge models and their available data, or to update model content.
For examples, see Using MCP Tools: Examples.
Troubleshooting
If GitHub Copilot CLI cannot connect to the Semaphore MCP Server, check the following:
| Problem | What to check |
|---|---|
| Server is not listed | Confirm that the configuration file is in the correct location and contains the Semaphore server definition. |
| Server is not connected | Verify the endpoint URL, network connectivity, and HTTP transport configuration. |
| OAuth authentication is requested | Use manual configuration with the authentication header appropriate for your deployment environment. |
| Authentication fails | Verify that the token is valid, has not expired, and is passed in the correct HTTP header. |
| Tools are unavailable | Confirm that the server is connected and that the authenticated user has the required permissions. |
| Configuration changes are not applied | Save the configuration file and restart GitHub Copilot CLI. |
Make sure the JSON configuration is valid.
For additional guidance, see Limitations and Troubleshooting.