Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
The settings for agents powered by the GitHub Copilot harness in Microsoft Copilot Studio let you configure how your agent behaves, processes content, and authenticates users. The Settings page organizes settings into categories.
Note
Features in this article are powered by the GitHub Copilot harness, which uses Copilot Credits for usage-based billing. Learn more in Overview of billing for agents powered by the GitHub Copilot harness.
Usage-based billing applies to using, building, testing, and evaluating agents. These actions might consume Copilot Credits. Learn more in Manage costs for agents powered by the GitHub Copilot harness.
Prerequisites
- Appropriate permissions to modify agent security settings
- For Microsoft authentication, access to Microsoft Entra ID configuration
Access agent settings
- Open or create your agent in Copilot Studio.
- Select the the three dots (…) in the agent designer toolbar.
- Select Settings. The Agent settings dialog opens, where you can browse and configure settings for your agent.
Configure settings
The Agent settings dialog is organized into four tabs.
Note
Changes to settings take effect after you save and publish the agent.
Agent details
The Agent details tab contains identity settings for your agent.
Important
You can only configure these settings before you save the agent for the first time. After the initial save, they become read-only and can't be changed.
In the Agent settings dialog, select the Agent details tab.
Configure the following settings:
- Schema name: The unique system name used to reference this agent. Leave empty to auto-generate from the agent name when you first save.
- Solution: The Power Platform solution this agent belongs to.
- Primary language: The language the agent prefers for its responses.
AI & behavior
The AI & behavior tab controls how the agent responds, reasons, and communicates.
In the Agent settings dialog, select the AI & behavior tab.
Configure the following settings:
- Allow other agents to connect: Toggle this setting to let other agents in your organization invoke this agent as a tool.
- Moderation level: Controls how strictly responses are filtered for unsafe content. Select Minimum, Low, Medium, High, or Maximum from the dropdown.
Safety & access
The Safety & access tab lets you manage authentication, channel-level protections, and how users rate responses.
- In the Agent settings dialog, select the Safety & access tab.
- Configure the following settings:
Authentication: How users authenticate when interacting with this agent (for example, Authenticate with Microsoft). Select one of the following methods from the Authentication dropdown:
- No authentication: Users can interact with your agent without signing in. Suitable for public-facing agents where user identity isn't required.
- Authenticate with Microsoft: Users sign in with their Microsoft account. Supported surfaces include Microsoft Entra ID, Teams, SharePoint, Power Apps, and Microsoft 365 Copilot. Recommended for internal agents that need to access user-specific data.
Web channel security: Required for compatibility with the Direct Line API on web channels. Select Web channel security.
- View your security secrets (Secret 1 and Secret 2). These secrets authenticate requests from your web channel.
- Select Copy to copy a secret for use in your web application.
- To regenerate a secret, select Regenerate next to the secret you want to replace.
User feedback: Toggle User feedback to show thumbs up and down on each response so users can rate quality.
Greeting & prompts
The Greeting & prompts tab configures the welcome experience for your agent.
In the Agent settings dialog, select the Greeting & prompts tab.
Configure the following settings:
- Greeting message: Enter the first message users see when they start a conversation.
- Suggested prompts: Prompts shown to users to help start a conversation. Select Add a suggested prompt to create prompts that help users start a conversation.