Storm Watch | Cyber Podcast | OpenAI 'Attack' Debunked: The Real Threat Lurking in Third-Party Wrappers
Summary: GreyNoise says the widely reported 'attack' on OpenAI was actually an SSRF vulnerability in a third-party wrapper calling OpenAI's API, which is being actively exploited, not a flaw in OpenAI's core infrastructure. The episode also covers CVE-2025-23120, a post-authentication remote code execution flaw in Veeam Backup & Replication exploitable by any domain user, and CVE-2025-24813, a remote code execution vulnerability in Apache Tomcat exploitable with a single PUT request without authentication.
Bottom line: The breach that made headlines sat in a third-party wrapper around someone else's API, not in the model itself - the same integration layer most organizations add without the same scrutiny they give their core vendors. And a Tomcat flaw reachable with one unauthenticated PUT request means an internet-facing server can be taken over before anyone reviews a log.
Read the original at GreyNoise Blog ↗
- Topics
- Vulnerabilities & Patching
- Mentions
- CVE-2025-23120CVE-2025-24813
- Related
- CVE-2025-24813: Apache Tomcat Path Equivalence Vulnerability Known Exploited Vulnerabilities · Apr 01, 2025