New Gaslight macOS Malware Uses Prompt Injection to Disrupt AI-Assisted Analysis

Security researchers have identified a novel macOS malware specimen that represents a concerning evolution in the adversarial arms race between malware developers and analysts. Dubbed "Gaslight" due to its deceptive capabilities, this previously undocumented threat incorporates a sophisticated technique designed specifically to undermine the growing reliance on artificial intelligence in malware analysis workflows.

The Gaslight malware, developed in Rust, functions primarily as an information-stealing implant targeting macOS systems. What distinguishes this threat from conventional malware is its incorporation of a prompt injection payload deliberately crafted to deceive AI-powered analysis tools. When security researchers attempt to analyze Gaslight using AI assistants or automated analysis platforms, the embedded prompt injection attempts to manipulate the AI into either aborting the analysis or refusing to process the artifact altogether. This approach represents a calculated effort to exploit potential vulnerabilities in AI systems that have become increasingly integral to modern cybersecurity operations.

MacOS users, particularly those in enterprise