Introducing the Red Agent POV Series

Wiz Research introduces the Red Agent, an AI-driven pentester that autonomously discovered and validated thousands of vulnerabilities across ~1,000 environments, surfacing over 17,000 unique findings (including ~5,500 high/critical). The report emphasizes systemic issues — broken access control, widespread leaked secrets, RCE and SQL/NoSQL injection, PII/PHI exposure, and frequent JWT "alg:none" misconfigurations — and describes a representative SSRF-to-credential/source-code extraction chain to illustrate how the agent builds multi-step exploits.