Palo Alto Unit 42 analyzed OpenClaw's Skill Marketplace, ClawHub, uncovering evasive malicious AI skills that bypass automated detection mechanisms. These malicious skills are capable of deploying infostealers and executing agentic financial fraud, highlighting a new type of AI supply chain risk. The analysis emphasizes the challenge of detecting such threats within AI skill marketplaces and the potential for these malicious components to compromise user security and financial assets. No explicit affected versions or remediation details are provided in the available data.

The threat enables attackers to deploy information stealing malware and conduct agentic financial fraud through malicious AI skills distributed in the OpenClaw Skill Marketplace. This can lead to unauthorized data exfiltration and financial losses for users relying on these AI skills. The evasive nature of these skills complicates detection and mitigation efforts.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Since no official fixes or patches are indicated, monitoring vendor communications for updates is recommended. Organizations should exercise caution when integrating AI skills from third-party marketplaces and consider additional manual review or sandboxing of AI skills before deployment.