reel-restyle

Pass

Audited by Gen Agent Trust Hub on Oct 10, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill orchestrates sibling Python scripts using subprocess calls in scripts/_restyle_common.py. It uses argument lists rather than shell strings, preventing shell injection.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted narration scripts via CLI or stdin and analysis data from JSON files. While it lacks explicit boundary markers, it enforces a manual review checkpoint for drafted storyboards before exercising its capabilities (file writes and subprocess orchestration). Sanitization is limited to whitespace normalization.
  • [DYNAMIC_EXECUTION]: Script paths for sibling tools are resolved at runtime relative to the skill or user directories. This dynamic resolution is standard for modular orchestration and follows the principle of least surprise for this tool type.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 10, 2026, 07:08 PM