advibly-pixar-style-ad
Pass
Audited by Gen Agent Trust Hub on Jul 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses ffmpeg and ffprobe for media analysis and mixing. These commands use hardcoded filenames and are executed to process the outputs of the skill's own generation tools, posing no identified security risk.
- [EXTERNAL_DOWNLOADS]: The skill uses curl to download media assets from URLs provided by its internal tools. These downloads are limited to media files required for the final mix and do not involve downloading or executing external scripts.
- [PROMPT_INJECTION]: The skill ingests brand and product data to generate model prompts. This surface for indirect prompt injection is mitigated by the use of rigid, structured prompt templates and sequential workflow steps. Ingestion points: Brand dossiers and product catalog data. Boundary markers: The skill uses sequential template blocks (Style, Subject, Action, etc.). Capability inventory: Local shell access for media processing and multiple AI generation tools. Sanitization: Not explicitly specified beyond structural prompt constraints.
Audit Metadata