veo3-1
Fail
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The troubleshooting section in
SKILL.mdincludes a command to install theuvpackage manager by downloading a shell script fromastral.shand piping it directly to the shell. Astral is a well-known provider of Python developer tools. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prompts to generate videos via an external API, which presents a surface for indirect prompt injection if the generated content or API response contains malicious instructions. 1. Ingestion points:
scripts/generate_video.py(via the--promptargument). 2. Boundary markers: None present. 3. Capability inventory: The skill can perform network requests to Grsai API endpoints and write generated video files to the local file system. 4. Sanitization: No sanitization or escaping is performed on the user-provided prompt before transmission.
Recommendations
- HIGH: Downloads and executes remote code from: https://astral.sh/uv/install.sh - DO NOT USE without thorough review
Audit Metadata