skills/rabbit-ivan/ivan-skills/sora/Gen Agent Trust Hub

sora

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a bundled Python script, scripts/sora.py, to perform video generation and management tasks. This script handles API interaction, job polling, and file management.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the openai Python package, which is fetched from the official Python Package Index (PyPI) when run with the recommended uv tool. This is a trusted dependency from a well-known service.
  • [PROMPT_INJECTION]: The skill processes untrusted user prompts as part of its core video generation functionality, creating an indirect prompt injection surface.
  • Ingestion points: Prompts are ingested through command-line arguments (--prompt) or external files (--prompt-file) as processed by scripts/sora.py.
  • Boundary markers: The skill implements a structured prompt augmentation template in scripts/sora.py to organize user inputs and define specific fields like 'Scene', 'Camera', and 'Action'.
  • Capability inventory: The script has capabilities for local file system read/write (for prompts, images, and video assets) and network communication with the OpenAI API.
  • Sanitization: Prompts are treated as data strings and transmitted via the official OpenAI SDK; they are not interpreted as executable commands by the script.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 11:00 AM