magic-hour-text-to-video
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references setup instructions, API documentation, and example projects hosted on the vendor's official GitHub organization and domain (magichourhq/magichour.ai). These are trusted resources used for configuring the Model Context Protocol (MCP) and understanding the API schema.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided descriptions ('ideas') to generate video prompts. While this is an ingestion point for untrusted data, the instructions guide the agent to translate these ideas into specific visual events for a structured API, minimizing the risk of instructions in the user input being executed as commands.
- Ingestion points: User-provided text descriptions for video generation.
- Boundary markers: The skill instructs the agent to define setup, change, and payoff, creating a structured interpretation of user input.
- Capability inventory: Calls to video creation and status tools, and file downloading of the resulting video.
- Sanitization: Not explicitly implemented, but the transformation into a visual shot direction serves as a natural barrier.
Audit Metadata