And Now Malware That Tells AI to Ignore It?

Researchers from Check Point analyzed a rudimentary malware prototype called "Skynet" found on VirusTotal that includes a hardcoded prompt intended to coerce AI analysis tools into declaring "NO MALWARE DETECTED." Although the prompt-injection was poorly implemented and failed against tested LLMs, the sample includes basic info-stealing and sandbox-evasion code; the report emphasizes this as an early sign of attackers experimenting with AI-targeted evasion rather than an actively dangerous, widespread campaign.