ai-avatar-video
Fail
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [REMOTE_CODE_EXECUTION]: User-supplied scripts and media paths are embedded directly into shell commands within JSON objects. A lack of proper sanitization or structured serialization creates a significant risk of command injection, allowing an attacker to execute arbitrary commands by including shell metacharacters in the input.
- [EXTERNAL_DOWNLOADS]: The skill directs users to install external CLI tools and dependencies using npx skills add without specifying pinned versions or integrity hashes. This introduces a supply chain risk as the external source content can change unexpectedly.
- [COMMAND_EXECUTION]: The skill requires extensive use of the belt CLI via bash, executing numerous external commands to perform media processing and remote model invocation.
- [DATA_EXFILTRATION]: Workflows involve transmitting user-provided images, audio, and text scripts to the external inference.sh platform for processing, which involves sending potentially sensitive data to a third-party service.
Recommendations
- AI detected serious security threats
Audit Metadata