agent-via-cli-skill

To automate GitHub Copilot agents from the command line, you launch them using specific CLI conventions for non-interactive, interactive, or delegate modes. This involves authenticating with Copilot and applying safe autopilot command patterns with explicit model selection and permission flags.

Yes, you can run Copilot custom agents in a CI pipeline using non-interactive launch modes. The CLI supports safe autopilot command patterns, explicit permission flags, and automated authentication to enable repeatable agent invocations across Linux, macOS, and Windows via WSL.

Available launch modes for scripting CLI agents include non-interactive, interactive, and delegate modes. These modes support various deployment and scripting workflows, allowing you to apply safe autopilot commands with explicit model selection and permission flags.

Yes, Copilot CLI supports both interactive and automated authentication for running custom agents. It provides installation and authentication guides tailored for interactive sessions and automated CI environments to ensure secure workflow orchestration.

To manage agent file scopes for repository-level automation, you use specific CLI file conventions to keep definitions with your code. This allows you to configure repository-level versus user-level scopes for custom agents directly from the command line.