animated-explainer
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The
scripts/build.pyutility usesimportlib.utilto dynamically load and execute the project'sspec.pyfile. This is a core architectural feature used to define the animation logic and timing for the video build process. - [INDIRECT_PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it executes arbitrary Python code from a file (
spec.py) that may be written or modified by the agent based on potentially untrusted external instructions. - Ingestion points: The
spec.pyfile located within the user's project directory (work/<reel>/spec.py). - Boundary markers: None. The content is executed directly as a Python module without delimiters or sandboxing.
- Capability inventory: The script has capabilities for subprocess execution, extensive file system access (read, write, delete via
pathlibandshutil), and dynamic module loading. - Sanitization: No validation or sanitization is performed on the content of the
spec.pyfile before it is executed. - [COMMAND_EXECUTION]: The
scripts/build.pyscript executes local Python scripts (lipsync.pyandpresenter.py) from the companioncartoon-characterskill usingsubprocess.run. These executions are performed using the current Python interpreter (sys.executable) with arguments derived from the local project state.
Audit Metadata