AI GitHub
Goose for extensible desktop and terminal agents
Connect a local agent interface to model providers and extensions for development tasks.
Key Takeaways
- Goose uses extensions to expose tools and offers a CLI and desktop experience.
- Choose the desktop or CLI workflow and connect a model and one extension.
- Run one read-only recipe against a disposable project.
- A locally running client can still use a remote model.
How it works
Goose uses extensions to expose tools and offers a CLI and desktop experience. Recipes can make a recurring task easier to repeat with a consistent set of instructions.
What you need
Choose the desktop or CLI workflow and connect a model and one extension. Review how the extension receives access to local data or network services.
What to try first
Run one read-only recipe against a disposable project. Inspect the enabled extensions and their available actions before turning the recipe into a workspace-editing task.
Run the recipe from a second checkout. Watch for hidden dependencies on a local path, cached session or credential present only on the first machine.
Where to look for failures
A locally running client can still use a remote model. Check where prompts and tool results go, and enable only the extensions the task needs.
Related Franklin pages
Compare how each agent exposes tools and where extensions execute. Start from an empty extension set and add one capability at a time so access stays understandable.
- Explore a terminal coding agent and the libraries behind it. Pi for customizing a coding-agent harness
- Give a coding agent access to editor context, workspace files and approved tools. Cline for agent-assisted development in an editor
For a library of development workflows that can complement a coding agent, see Franklin’s Superpowers profile. Superpowers
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗