scenario-sprite-animation
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes character descriptions and animation notes from a configuration file and interpolates them into natural language prompts for generative models without sanitization.
- Ingestion points: Fields like 'character', 'notes', and 'walk_motion' in the cast JSON configuration file are read by 'cycle_prompts.py'.
- Boundary markers: Absent; data is merged directly into prompt strings without delimiters.
- Capability inventory: The skill calls 'model_run' to generate images and video, and executes 'ffmpeg' via subprocess calls in 'cycle_sheets.py'.
- Sanitization: No escaping or validation is performed on the user-provided strings before they are used in generation prompts.
- [COMMAND_EXECUTION]: The script 'scripts/cycle_sheets.py' executes the 'ffmpeg' and 'ffprobe' binaries to handle video frame extraction and metadata retrieval.
- Evidence: The script uses 'subprocess.run' with a list of arguments to call the external binaries 'ffmpeg' and 'ffprobe'.
- Context: The implementation avoids shell expansion by passing arguments as a list, which mitigates typical command injection risks, and serves the legitimate primary purpose of the skill.
Audit Metadata