Closed jwmatthews closed 2 weeks ago
Extended what was started in #312 in an experiment to redo run_demo.py without solved incidents.
Idea is to move to the generate fix logic to client side, then add an agent in to take the result and vet it and work with a LLM to improve it.
At the moment this PR has the capability of run_demo.py, it does not yet have an agent in it.
Extended what was started in #312 in an experiment to redo run_demo.py without solved incidents.
Idea is to move to the generate fix logic to client side, then add an agent in to take the result and vet it and work with a LLM to improve it.
At the moment this PR has the capability of run_demo.py, it does not yet have an agent in it.