CertiK Skill Scanner: Antivirus for AI Agents Launched

CertiK has launched CertiK Skill Scanner, a new security solution for AI Agents and third-party AI Skills. This product aims to secure the rapidly growing AI ecosystem. The scanner is designed for AI Skill marketplaces, enterprises, developers, and eventually everyday users. It addresses concerns about malicious behavior, unauthorized data access, and autonomous execution risks in AI environments. CertiK's CEO, Ronghui Gu, states that the security model for third-party Skills is critically important as AI Agents integrate into financial systems and daily interactions. What's interesting is that this tool evaluates risks during actual execution, including scenarios with fund calls and financial transactions. It works across both Web3 and traditional Web2 ecosystems. The scanner generates a scored assessment from 0 to 100 with "pass," "warn," or "fail" verdicts. CertiK says the system achieves up to 90.5% precision in identifying security risks. This matters because it provides a new layer of security for anyone interacting with AI tools, aiming to build trust and prevent potential harm.