p-video-2
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the installation of prerequisite skills from the vendor's GitHub organization 'PrunaAI' using the npx command.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided text prompts and external media URLs for video generation. This constitutes a data ingestion surface for indirect instructions. * Ingestion points:
SKILL.md(viaprompt,image, andaudioinput fields). * Boundary markers: Present (includes explicit 'Fidelity check' and 'Prompt craft' instructions for the agent). * Capability inventory: Performs HTTP POST requests toapi.pruna.aifor prediction tasks. * Sanitization: Absent (rely on model-side API safety filters). - [COMMAND_EXECUTION]: The skill provides example
curlcommands for the user to interact with the Pruna AI prediction API.
Audit Metadata