Hacker News new | ask | show | jobs
by holoduke 64 days ago
With LLM tool use potentially every cat action could be a prompt injection