ChxisB/Talon — reverse-engineered prompt

Build me a local terminal AI coding assistant called Talon that feels fast and polished for everyday coding work. I want a full screen terminal interface with chat, a file tree, session history, model selection, and a nice workflow for jumping between code and conversation.

It should work with multiple AI providers and also local models, so I can plug in whichever API key I have. The assistant should be able to read and edit files, run shell commands with permission prompts, search through the project, fetch web pages, and use language aware context when possible. I also want screenshot capture and image analysis, including support for a local vision model if Ollama is running.

Please include persistent sessions, smart context handling for long chats, and a simple skills or sub agent setup if it fits naturally. I also want a web dashboard where I can see status, tasks, memory, settings, and live activity.

Make it runnable on macOS or Linux, and easy to start with Docker too. If needed, look up current docs online and wire up a smooth first run setup for provider keys and config.