@technicallymatthq: The OpenAI Hugging Face security incident was caught by monitoring, not by the sandbox. What that means if you run AI agents. Comment what your own agent logs would actually show you. #aisecurity #aiagents #openai #huggingface
Are openAI being charged with cybercrimes? Seems only fair As a person would be for sure.. so surely the company that did it should be charged too…
2026-07-22 21:55:22
0
𝙕𝘼𝙄𝙉.🚩 :
Love this — I'm building in the same space. It's
called AgentWall: a security layer that sits between
AI agents and company systems, blocking prompt
injection and other attacks in real time before they
can do damage. Focused on UK fintech first.
Great to see more people building for agent security 👏
2026-07-22 23:17:54
0
mantis_tobogan :
Which specific zero-day vulnerabilities were exploited? Were they vulnerabilities in the third-party reverse proxy/package cache used by OpenAI's sandbox, or were they vulnerabilities in Hugging Face's own internally developed software or infrastructure?
2026-07-23 07:29:17
0
To see more videos from user @technicallymatthq, please go to the Tikwm
homepage.