automotive-vehicle-security
Audited by Socket on Sep 19, 2026
2 alerts found:
Securityx2SUSPICIOUS: the skill is internally consistent as an automotive offensive security guide and its installs/data flows are mostly normal, but it gives an AI agent explicit exploit and real-world attack procedures against vehicles. This is not deceptive malware, yet it is a high-risk offensive capability skill with moderate supply-chain risk and significant physical-world misuse potential.
The fragment contains explicit CAN attack and diagnostic-testing proof of concepts, including frame injection, intentional bus denial of service, and VIN querying. These actions are high impact on a real vehicle network and should remain confined to authorized vcan or bench environments. No evidence of malware, persistence, credential theft, remote exfiltration, or obfuscated malicious payloads is present. The content presents operational security risk if misused, but it does not itself indicate supply-chain malware.