elevenlabs-dubbing
Pass
Audited by Gen Agent Trust Hub on Jun 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: References installation instructions from the official GitHub repository of the inference-sh organization (raw.githubusercontent.com/inference-sh/skills).
- [COMMAND_EXECUTION]: Utilizes the
beltCLI tool to execute dubbing, speech-to-text, and captioning applications. This is the core intended functionality of the skill and is restricted to the specific CLI tool via the allowed-tools configuration. - [DATA_EXPOSURE_AND_EXFILTRATION]: Provides instructions for authenticating with the service via
belt login. This is a standard procedure for CLI-based services and does not involve hardcoded credentials. - [INDIRECT_PROMPT_INJECTION]: The skill creates a data processing pipeline where external media files are ingested and their outputs (transcripts) are passed to subsequent tools.
- Ingestion points: Processes external media files via the
audioandvideo_urlparameters in SKILL.md. - Boundary markers: Uses structured JSON inputs which provide a degree of separation between instructions and data.
- Capability inventory: Executes remote applications via the
beltCLI (SKILL.md). - Sanitization: Relies on the underlying platform and CLI tools to handle data safely.
Audit Metadata