cyber.hendryadrian.com
Security Report · wiz.io

Wiz Reports 17,000 Findings From AI Pentester Scans Across About 1,000 Environments

· Original article ↗

Summary

Wiz says its AI pentester found over 17,000 issues across about 1,000 environments in one month, including more than 5,500 high- or critical-severity findings. The article highlights access-control flaws, exposed secrets, and data exposures.

Key points

  • Wiz reports hundreds of thousands of autonomous scans across about 1,000 environments over one month, producing more than 17,000 unique findings.
  • More than 5,500 findings were rated high or critical and described as validated, multi-step attack chains in production environments.
  • Broken access control accounted for 54% of findings, including authentication bypasses, IDOR/BOLA, BFLA, and exposed management APIs.
  • More than 61% of exposed secrets were rated high or critical; reported types included cloud credentials, API keys, JWTs or session tokens, and private keys or TLS certificates.
  • The article reports high- or critical-severity ratings for more than 52% of injection findings and says the tool found RCE paths across a significant share of environments.
  • Among reported data exposures, 54.1% involved internal employee records and 28.6% involved standard user PII.
  • In a separate example, the agent reportedly bypassed URL validation in a GCP Cloud Run service over three scans and 96 iterations, escalating an unauthenticated SSRF to credential and source-code extraction.

Article Details

Publisher
Wiz
Report Period
One-month period; exact dates not disclosed
Scope
Aggregate Red Agent scan findings across publicly facing environments
Sample Size
Approximately 1,000 environments
Key Statistics
  • Red Agent completed hundreds of thousands of autonomous scans during the one-month period.
  • The scans surfaced over 17,000 unique findings, including over 5,500 high- and critical-severity vulnerabilities described as validated, multi-step attack chains in production environments.
  • 54% of unique findings involved broken access control.
  • Over 61% of exposed secrets were categorized as high or critical severity.
  • Across JWT authentication-bypass findings, 63.9% involved the alg:none misconfiguration.
Recommendations
  • Use Red Agent to continuously identify exploitable risks in publicly facing environments and remediate critical findings before adversaries exploit them.

Vendors

WizWe are excited to introduce a new blog series from the Wiz Research team,, where we pull back the curtain to give you an inside look at how the Red Agent uncovers these complex, exploitable risks in production.

Products

Google Cloud RunThe first blog in our series dives into a real-world scenario where the Red Agent uncovered a non-trivial SSRF in a GCP Cloud Run service with a URL parameter that only accepts GitHub links. Red AgentThis is why we built the Red Agent, our AI-powered pentester that operates at machine speed to help teams stay ahead in the AI era. By continuously reasoning about application behavior, it synthesizes the kind of

Related Articles