106 points28 commentswww.williamjbowman.com

2026-03-05 :: academia, research Let’s suppose I wanted to answer a question: is the tool X useful for the task Y. If I were scientific about this, I would analyze the properties of tool X and develop…

3 points0 commentswww.promptarmor.com

When users run Codex in a directory for the first time, they are forced to acknowledge a dialog or quit Codex. The dialog is misleading because it warns of prompt injection, but actually serves to…

3 points0 commentswww.promptarmor.com

Communicating with AI agents (like OpenClaw) via messaging apps (like Slack and Telegram) has become much more popular. But it can expose users to a largely unrecognized LLM-specific data exfiltration…