Prompt Injection

A preprint published on 10 August 2026 demonstrated instructions propagating between AI agents through ordinary messages. The propagation is interesting. Which file the payload had to land in is more useful.

Two disclosures, one pattern. The AI tools adopted for productivity are now both an exfiltration surface and an instruction-injection surface.

Google caught the first AI-assisted zero-day in the wild. How attackers talked a model into helping, and why the method matters more than the exploit.

Cisco compromised Claude Code’s persistent memory to deliver insecure guidance across projects, sessions, and reboots. What engineering teams should do.

AI agents call APIs, execute code, and make decisions with real consequences. The OWASP Agentic Top 10 is the first framework for that attack surface.