make-a-video
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the shell to execute various utility commands via npx, including 'hyperframes' for framework operations (init, lint, preview, render, transcribe, tts) and 'serve' for hosting local video previews. It also uses ffmpeg for frame extraction and visual verification steps.- [EXTERNAL_DOWNLOADS]: The skill fetches typography assets from an official web fonts service and downloads framework-specific components using the 'npx hyperframes add' command.- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted external data, such as the 'MOTION_PHILOSOPHY.md' file from the workspace root and user-provided outlines or scripts, to generate core project artifacts like the brief and storyboard.
- Ingestion points: Accesses 'MOTION_PHILOSOPHY.md' and various project asset directories to gather context.
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the prompts that process external data.
- Capability inventory: The agent has access to file system operations (creating directories and writing files) and shell command execution (npx, ffmpeg).
- Sanitization: No evidence of input validation or content sanitization is present for the data read from the workspace.
Audit Metadata