ESET Reports GuardBreaker Technique Used by UAC-0099 Against Ukrainian Target

According to ESET, the Russia-aligned threat actor UAC-0099 deployed a new technique codenamed GuardBreaker against a target in Ukraine. The stated objective was to interfere with artificial intelligence-assisted analysis by deliberately triggering large language model (LLM) safety mechanisms. ESET shared details of the technique on its official X account, describing the method as an attempt to disrupt AI-assisted security workflows. The Hacker News reported on the disclosure, noting that the technique's primary goal was to prevent or mislead AI analysis. As of the report, no specific technical details of the GuardBreaker implementation have been fully disclosed, and the actual effectiveness of the method against deployed AI systems is unclear. The incident highlights the growing intersection between cyber threat operations and artificial intelligence infrastructure. ESET and other researchers advise that AI outputs in security-critical contexts should be subject to human oversight and robust validation.