Red-team LLM reasoning and agent actions (honest scoring, local-first)github.com4 pointsby rudrasatani0 commentsSharePost on XLinkedInCopy post