Limitation of Bash tools in LLM Agents?

«I am trying to see how successful bash tools are in LLMs such as Claude etc. The research I am conducting is specifically in reverse engineering malware samples. There might be encrypted or obfuscated parts of the code (...»