creative-video
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows secure coding practices, including the use of environment variables for API keys and SHA256 checksum verification for external dependencies and models downloaded from trusted sources like Hugging Face and NuGet.
- [DATA_EXPOSURE]: The skill contains a defensive mechanism in
scripts/source/snapshot.mjsthat automatically identifies and excludes sensitive files (e.g., .env, .pem, .key) and common secret patterns from being uploaded to the agent's context when a local repository is used as a source. - [COMMAND_EXECUTION]: Standard media tools such as
ffmpegandffprobeare used for audio/video processing. These commands are executed with internally defined parameters to fulfill the skill's primary objective. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data from the web and various document formats. It mitigates injection risks by using specialized sub-agents to extract facts and excerpts into structured formats before they are used in the final script generation.
- [EXTERNAL_DOWNLOADS]: All external downloads originate from well-known services (Google, ElevenLabs, Hugging Face, Fontsource, Openverse) and are strictly limited to assets necessary for video creation, such as fonts, audio clips, and AI models.
Audit Metadata