elevenlabs-dialogue
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references and points to installation documentation hosted on the inference-sh organization's public GitHub repository.
- [COMMAND_EXECUTION]: Utilizes the belt CLI tool to perform audio generation tasks and handle authentication. The execution environment is restricted to only allow belt commands through the allowed-tools configuration, adhering to the principle of least privilege.
- [INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: The skill ingests dialogue text through the segments array in the JSON input for the belt app run command. 2. Boundary markers: Uses structured JSON for data ingestion, though it lacks explicit delimiters to prevent the agent from obeying instructions embedded within the text segments. 3. Capability inventory: The Bash tool is enabled specifically for the belt CLI, which performs network operations to reach the inference.sh API. 4. Sanitization: No specific sanitization or filtering of the input text is performed before it is sent to the external audio synthesis engine.
Audit Metadata