A flaw in DeepSeek Harness, DeepSeek's opensource tool for running AI coding agents on a developer's machine, let a sandboxed agent turn off its own sandbox with a single command. The tool runs an agent's commands inside an operatingsystem sandbox, so that an agent working on untrusted files cannot write outside its workspace. The agent could remove that limit by calling the tool's own web.📖 Read more.
🔗 Via "The Hacker News"
----------
👁️ Seen on @cibsecurity