speech-to-text
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the official ElevenLabs CLI and SDKs from trusted package registries including npm, PyPI, and Homebrew. It also references the official ElevenLabs GitHub repository for Windows installation via Scoop.
- [REMOTE_CODE_EXECUTION]: Includes a shell installer script hosted on the official ElevenLabs GitHub releases page (https://github.com/elevenlabs/cli/releases/latest/download/elevenlabs-cli-installer.sh). This is the standard, vendor-supported installation method for the CLI tool.
- [COMMAND_EXECUTION]: Provides examples for executing the
elevenlabsCLI for audio-to-text conversion and authentication, which is the primary intended functionality of the skill. - [INDIRECT_PROMPT_INJECTION]: As a transcription tool, the skill processes untrusted audio and video content which could contain embedded instructions intended to influence the agent.
- Ingestion points:
client.speech_to_text.convert,connection.stream_url, and the CLI--file/--source-urlinputs. - Boundary markers: The tool returns results in a structured JSON format, reducing the risk of schema confusion.
- Capability inventory: The skill is scoped to transcription, diarization, and entity detection.
- Sanitization: Not explicitly described in the documentation.
Audit Metadata