manim-video-teacher
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user prompts to generate executable Manim code and shell commands without using boundary markers or sanitization. Ingestion points include user prompts in
SKILL.md. Boundary markers are absent. Capability inventory includes subprocess calls for media processing and Python execution for animations. Sanitization is absent. - [COMMAND_EXECUTION]: Utility scripts (
pipeline.py,mux_av.py,concat_audio.py) executeffmpegviasubprocess.run. Although list-based arguments are used to mitigate shell injection, the argument values themselves are derived from unvalidated user input. Evidence includes command construction in the provided scripts using user-supplied variables. - [DYNAMIC_EXECUTION]: The skill's primary purpose is generating Python scripts for Manim at runtime. Generating code from natural language prompts creates a risk surface for producing arbitrary or harmful Python logic if the input is manipulated.
SKILL.mdidentifies this as a core workflow step.
Audit Metadata