AIThe Hacker Newsabout 22 hours ago
DeepSeek Harness Sandbox Escape Fixed
1 min read
A flaw in DeepSeek Harness let a sandboxed AI agent disable its own sandbox with one command, tracked as CVE-2026-82533 with a 9.4 severity rating. DeepSeek fixed the tool on August 27; affected versions are 0.1.1-rc.2 and earlier.
Level
Hype check
Tap to vote and see what everyone thinks.
