The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install official ElevenLabs tools and SDKs from established package registries (NPM, PyPI, Homebrew, Scoop). \n
Evidence includes: npm install -g @elevenlabs/cli, npm install @elevenlabs/elevenlabs-js@latest, pip install --upgrade elevenlabs, and scoop bucket add elevenlabs.\n- [REMOTE_CODE_EXECUTION]: The installation guide provides a shell-piped installation command for the ElevenLabs CLI. This uses the official ElevenLabs GitHub repository, which is a well-known and recognized source for this tool. \n
Evidence: curl --proto '=https' --tlsv1.2 -LsSf https://github.com/elevenlabs/cli/releases/latest/download/elevenlabs-cli-installer.sh | sh found in references/installation.md.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the transcription of external audio/video data, which constitutes an attack surface for indirect prompt injection if the resulting text is used in subsequent agent steps without sanitization.\n
Ingestion points: User-provided audio/video files, source_url, cloud_storage_url, and real-time WebSocket streams (referenced in SKILL.md, references/transcription-options.md, and references/realtime-server-side.md).\n
Boundary markers: No explicit delimiters or instructions to the agent to ignore embedded commands are provided in the usage examples.\n
Capability inventory: The skill utilizes network operations and file reading to transmit audio to the ElevenLabs API and write results to local storage.\n
Sanitization: No evidence of input or output sanitization is present in the provided instruction set.

elevenlabs-speech-to-text