ylxdzsw

HackXKCD please

Probably get a PR in Hong Kong and hope to be automatically "promoted" to Chinese in 20 years?

sandboxing is indeed an advantage (can be an important one!), but

1. bash can also express concurrency easily, with sync and async (using standard & syntax) support for each command, and standard cancellation (kill, though crude). 2. "way fewer instructions": It requires no instruction for agents to use bash either, except for merely listing the custom commands (view_image, apply_patch, etc.). Also, bash has standard progressive disclosure mechanism (--help) that models will automatically use with no instruction.

I'm not sure why almost all codemode implementations choose Javascript. I prototyped an agent[1] to use bash as the language for codemode, which in my opinion worked equally well and requires no teaching (there is literally 0 prompt to teach the LLM about codemode. A tool named "bash" is enough to have them know the usage).

[1] https://github.com/ylxdzsw/mu

Better, train it to attack when receiving the "sandbox lifted" / "full access mode" note in Codex