The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text prompts and uploaded audio files, which are then passed to the ElevenLabs API. It also handles structured API responses that include content suggestions, representing a potential surface for indirect instructions.
Ingestion points: User prompt parameter in SKILL.md and file upload in references/api_reference.md.
Boundary markers: The skill does not provide specific instructions to the agent to use delimiters or ignore instructions embedded within user data.
Capability inventory: The skill performs network requests to an external API and writes generated audio to the local file system (output.mp3).
Sanitization: Input validation and safety filtering are handled by the remote ElevenLabs service; the skill itself does not perform pre-processing or sanitization of the input.
[EXTERNAL_DOWNLOADS]: The skill documentation references the installation of official libraries from established registries.
Evidence: Mentions npm install @elevenlabs/elevenlabs-js and pip install elevenlabs in references/installation.md.
[COMMAND_EXECUTION]: The skill provides examples for using command-line tools to interact with the service endpoint.
Evidence: cURL commands targeting https://api.elevenlabs.io/v1/music are included for testing and implementation purposes in SKILL.md.

music