LLM hacker
Almog Baku builds skills that let an AI agent work through live debugging sessions, setting breakpoints, stepping through execution, inspecting variable state and tracing call stacks to find the root cause of crashes and wrong output. The listed work centers on code diagnosis, a practical and hands-on area of agent tooling. Almog's GitHub bio describes them as an LLM hacker.
Debug Skill
1 skills · 336 stars · MIT
Debugging Code Interactively debug source code — set breakpoints, step through execution line by line, inspect l…