Skip to content

Back to the catalog

レッドエージェント POVシリーズのご紹介
wiz.ioarticle

レッドエージェント POVシリーズのご紹介

Wiz Research introduces the Red Agent AI-powered pentester and its new blog series documenting real bug classes it discovers autonomously in production. The launch post shares aggregate stats from a one-month run across roughly 1,000 environments, breaking down dominant failure modes: broken access control, leaked secrets, systemic RCE, SQL injection, PII/PHI exposure, and JWT alg:none bypasses. It teases the series first deep-dive: a GCP Cloud Run SSRF the agent found by iteratively bypassing a GitHub-URL-only validation filter over 96 iterations, escalating to unauthenticated credential and source-code extraction.

TL;DR Wiz launches a blog series on its AI pentesting agent Red Agent, sharing aggregate vulnerability stats from autonomous scans of ~1,000 environments and teasing a GCP Cloud Run SSRF case study.
Open original ↗

Organisations named here

  • JWT alg:none
  • Wiz Research

Technologies named here

  • Red Agent
  • GCP Cloud Run service
  • Wiz
  • GCP

Techniques named here

  • SSRF
  • Broken access control
  • Leaked secrets exposure
  • Remote code execution
  • SQL/NoSQL injection
  • PII/PHI data exposure
  • JWT alg:none authentication bypass

Related in AI Security