An LLM-powered Slackbot
TODO:
- Proper logging
- Stable Diffusion
- Improved determination of when to reply
- Fix whitespace output in reply determination
- Separate context by channel?
- Smarter message retention (by token count)
- Refactor for maintainability / extensibility
- Better error handling
- LLaVA support
- DevEx (tooling improvements)
- Tests