The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation guides the installation of official ElevenLabs packages and tools from standard registries. Other well-known supporting libraries are also referenced.
Official CLI: @elevenlabs/cli (NPM), Homebrew tap elevenlabs/tap, and Scoop bucket elevenlabs.
Official SDKs: @elevenlabs/elevenlabs-js, @elevenlabs/client, and @elevenlabs/react (Node.js/NPM), as well as elevenlabs (Python/PyPI).
Supporting libraries: python-dotenv, pydub, and dotenv are referenced for environment management and audio processing.
[REMOTE_CODE_EXECUTION]: The installation guide provides a shell command to download and execute an installer script for the ElevenLabs CLI from the vendor's official GitHub repository.
Evidence: curl --proto '=https' --tlsv1.2 -LsSf https://github.com/elevenlabs/cli/releases/latest/download/elevenlabs-cli-installer.sh | sh in references/installation.md.
This script is the official standalone installation method provided by ElevenLabs.
[INDIRECT_PROMPT_INJECTION]: The skill enables transcription of audio and video content from external, user-provided URLs, which introduces a potential attack surface for indirect prompt injection via the resulting transcript text.
Ingestion points: source_url, cloud_storage_url, and stream_url parameters documented in references/transcription-options.md and references/realtime-server-side.md.
Boundary markers: The skill instructions do not specify delimiters or boundary markers for the transcribed text output provided to the agent.
Capability inventory: The skill transcribes external media via the ElevenLabs API, which is then processed by the agent; the SDKs and CLI tools facilitate these network operations.
Sanitization: There is no mention of internal sanitization or validation of the transcribed text content within the provided documentation.

speech-to-text