google-veo

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches installation instructions from the inference.sh official GitHub repository.
  • [COMMAND_EXECUTION]: Utilizes the 'belt' CLI tool to interact with video generation models. The skill's configuration limits the agent's command-line access to this specific tool.
  • [INDIRECT_PROMPT_INJECTION]: The skill takes user-defined text prompts to generate video content, which is a standard surface for data processing.
    1. Ingestion points: User-provided prompts passed to the 'belt' CLI via bash commands in SKILL.md.
    1. Boundary markers: None present in the skill instructions.
    1. Capability inventory: Shell execution restricted to the 'belt' CLI via the Bash tool.
    1. Sanitization: None mentioned; the skill relies on the inference.sh platform's internal content filters.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:54 PM
Security Audit — agent-trust-hub — google-veo