What happened
On Oct 21 TrendAI (Trend Micro's AI security unit) announced two Vision One capabilities: private preview of Autonomous Vulnerability Discovery (AESIR agentic exploit-remediation engine, Anthropic-model-based, runs on Amazon Bedrock inside the customer AWS environment, zero-day finding validation by ZDI) and expanded Claude Compliance API integration bringing Claude Code and Cowork session logs (prompts, responses, tool calls per verified user) into Vision One for investigation and custom detection.
Why it matters
Autonomous source-code vulnerability discovery with human (ZDI) validation is a meaningful anti-hallucination control versus fully autonomous scanning peers, and native Claude Code/Cowork telemetry closes an agent-observability gap that pure network products miss.
Applicability
AWS-based enterprises running TrendAI Vision One should request private preview access for vuln discovery; security teams using Claude Enterprise should plan around Vision One ingesting Claude Code/Cowork activity for agent-behavior detections.