Red-teaming AI agent monitors with an LLM that fakes tool outputs redwoodresearch.org 2 points by sgiannopoulos 18 days ago · 2 comments Reader PiP Save No comments yet.