video-remake
Warn
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes shell commands and CLI tools such as
yt-dlpfor downloading video content from external sources. It also references asandboxtool for processing. - [DYNAMIC_EXECUTION]: Instructions authorize the agent to generate and execute its own code for tasks like video editing, compositing, and motion graphics when a shell or sandbox environment is available.
- [EXTERNAL_DOWNLOADS]: Fetches video data from various well-known third-party platforms including YouTube, TikTok, and Instagram to serve as reference material for the remaking process.
- [INDIRECT_PROMPT_INJECTION]: The skill processes and analyzes untrusted external content (videos and links). It relies on host content moderation to handle potential safety violations. Ingestion points: User-provided video files or URLs. Boundary markers: Structured JSON plan and placeholders. Capability inventory: Shell access, yt-dlp, and sandbox tools. Sanitization: Adherence to host-provided content-review errors.
Audit Metadata