The Hidden Security Risks of AI Coding Agents
19 May 2026
Agentic systems introduce heightened security risks through text-based interactions enabling malicious intent encoding, sensitive data access, untrusted inputs, and external system communication, requiring mitigation via SCA, restricted agent access, dynamic analysis, and balancing security with productivity through transparency and adapted security frameworks.
Open episode




![[Outliers] The Hyundai Founder Who Put a Country on His Back thumbnail](https://megaphone.imgix.net/podcasts/f6f18d8e-52e7-11f1-be70-f3a00b01d39b/image/f0cb7c360107f68efff87ed148174a79.png?ixlib=rails-4.3.1&max-w=3000&max-h=3000&fit=crop&auto=format,compress)



