Questions on the system

[TomLucidor]

1. Can it mix multiple LLMs?
2. Can it self-improve and learn from mistakes?
3. Can it create playbooks, and then share them ala "borg"?
4. Can this system do research instead of SWE?

[blackms]

1. yes
2. it's on roadmap
3. not planned but I'm open to hear ideas. currently it's support workflows.
4. this really depends on what you mean with research. it depends on LLM models, not the software itself.

[TomLucidor]

@blackms LLMs can read whatever webpages or papers, but the software determines how things are read. As for borg... https://github.com/aplaceforallmystuff/claude-agent-borg