CertiK Launches Skill Scanner Security Tool for AI Agents in Web3 and Web2

According to CertiK’s May 27 announcement, Skill Scanner detects security risks in third-party AI Skills with 90.5% accuracy and has been integrated into Pieverse’s AI Agent Skill store.

Summary

CertiK launched CertiK Skill Scanner, an AI Agent security product for Web3 and Web2 environments, and said in its May 27 announcement that the tool achieves 90.5% risk detection accuracy. The product is designed to scan third-party AI Skills for malicious behavior, data leakage, unauthorized access, and execution risks before a skill is used. CertiK also said Pieverse has integrated Skill Scanner into its AI Agent Skill store, adding an early deployment example for the product.

Terms & Concepts
  • Web3: A blockchain-based internet model centered on decentralized applications, digital assets, and user-controlled data.
  • AI Agent: Software that can perform tasks autonomously, often by making decisions and executing actions based on inputs or goals.
  • AI Skills: Third-party modules or capabilities used by AI Agents to perform specific tasks, which can be screened for security risks before use.