Dataset: AI agent security failures, 1000 incidents classified

A dataset has been compiled that classifies 1000 incidents of AI agent security failures. The dataset is available on Hugging Face and a corresponding discussion can be found on Hacker News. The dataset is titled ai-agent-security-incidents and was posted by gemmozero. It has garnered 2 points but no comments have been made on the topic yet. The resource is provided for those interested in exploring AI agent security incidents.