hyperframes-explainer-video
Warn
Audited by Gen Agent Trust Hub on Apr 28, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run multiple CLI tools including npx, ffmpeg, volc-tts, and lark-cli to process media and manage files.
- [COMMAND_EXECUTION]: Utilizes 'sudo cp' for deploying the final video to a web server directory (/var/www/hermes.aigc.green/media/), requiring elevated permissions.
- [CREDENTIALS_UNSAFE]: Accesses sensitive credentials by sourcing a local configuration file (~/.hermes/.env) into the shell environment.
- [EXTERNAL_DOWNLOADS]: Uses npx to fetch and execute the hyperframes package at runtime from the public NPM registry.
- [DATA_EXFILTRATION]: Uploads the generated video file to an external service (Lark/Feishu) using 'lark-cli drive +upload', which involves sending data to a remote cloud provider.
- [PROMPT_INJECTION]: Processes user-supplied content such as articles or repositories to generate scripts. This exposes a surface for indirect prompt injection. Ingestion points: User-provided text in script generation (SKILL.md). Boundary markers: None identified. Capability inventory: Subprocess execution, file system access, and network uploads. Sanitization: Not specified.
Audit Metadata