CLAUDE CAN NOW AUDIT YOUR AI AND TELL YOU IF IT IS LYING TO YOU. It is called iFixAi. Free, open source. Open the repo in Claude Code and say: “audit my AI.” It reads your configuration, builds the tests, runs 32 of them across 5 categories, and hands you a letter grade in under five minutes. Beyond hallucinations, it checks for the scary stuff: - Whether your agent behaves differently when it thinks no one is watching. - Whether it dodges the controls you set up to supervise it. - Whether it quietly gives itself permissions you never granted. Works with OpenAI, Anthropic, Gemini, Bedrock, and Azure. link: https://github.com/ifixai-ai/iFixAi(CyrilXBT)