alterlab-genai-dubbing-specialist
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides legitimate functionality for video dubbing and localization using the ElevenLabs platform. All instructions and suggested workflows are consistent with the stated purpose.
- [INDIRECT_PROMPT_INJECTION]: The skill includes an attack surface for indirect prompt injection through its core data ingestion functions.
- Ingestion points: External content imported via YouTube, Vimeo, and X URLs, as well as uploaded media files (MP4, MOV, MP3, WAV).
- Boundary markers: None explicitly specified for the ingested content or generated transcripts.
- Capability inventory: The agent can read and write files to the local project directory and perform web searches for platform updates.
- Sanitization: No specific sanitization or filtering logic is described for the content extracted from external URLs or audio files. This surface is considered a low-risk inherent characteristic of the skill's primary dubbing function.
Audit Metadata