google-veo

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill is a plain Veo video-generation guide, but its install pointers (a belt-sh/cli skill from an organization other than the inference-sh one it documents, and an unvetted raw-URL install script for the CLI) are ambiguous, sketchy, and unexplained, so I flag them as medium risk without evidence of malicious intent.