Moreover, I added some Ruff rules in the

sdk

to reduce the verbosity of the code. Just a proposal: github.com/OpenHands/software-agent-sdk/pull/4226 Let me know what do you think. cc @calm-artist-7206

I’m not sure I like the reduction of:

if x:
  return True
else:
  return False

to

if x:
  return True
return False

The first seems clearer to me

The way I’d think of it, is applying the LLM-lens, or to put it differently, what kind of rules are more likely to lead to LLM understanding better the resulting code. I do like the explicit rules better for this reason too

My idea was that less code less context. I don’t think there are fixed code rules to make LLMs work better. Moreover that we are model agnostic.

Of course , that’s the beauty of it, isn’t it? 😅