How Claude Code escapes its own denylist and sandbox | Ona - AI software engineers
Today we're releasing Veto in early access, our content-addressable kernel enforcement engine.
In the last ten days: a single person used Claude to breach Mexican government agencies. Cline's own AI-powered triage workflow was compromised via prompt injection. A new Shai-Hulud variant started injecting rogue MCP servers into developer AI tools.
In 2020 I gave a talk called "Bypass Falco" where I showed an audience how to break the CNCF runtime security tool I helped create. Symlinks, renamed bin...
Read more at ona.com