TrendAI Teams With Anthropic to Find AI Platform Vulnerabilities, Discovers CVSS 9.8 Flaw in Nvidia Isaac GR00T

According to Trend Micro, TrendAI is collaborating with Anthropic to use Claude Opus 4.7 for security research aimed at identifying exploitable software flaws and ranking them by risk. TrendAI has joined Anthropic's Cyber Verification Program, which grants approved groups access to frontier AI models for defensive cybersecurity work.

TrendAI's AESIR research platform, launched in 2025, uses Claude Opus 4.7 to assess whether vulnerabilities are reachable and exploitable. The platform identified a remote code execution bug with a CVSS score of 9.8 in Nvidia's Isaac GR00T robotics platform, tied to a deserialization error in the TorchSerializer class. TrendAI has also worked with the Zero Day Initiative on patches for vulnerabilities affecting AI products from Nvidia and Tencent.