The Agent Skills Directory

[DATA_EXFILTRATION]: The helper script reads ELEVENLABS_API_KEY from the environment or local .env files and transmits it to api.elevenlabs.io. This is the intended behavior for an API integration and targets a well-known service.
[INDIRECT_PROMPT_INJECTION]: The skill processes text from external files, standard input, or command-line arguments to generate speech audio.
Ingestion points: The generate_voice.py script accepts text input through the --text argument, --text-file path, or via stdin.
Boundary markers: The skill does not implement specific delimiters or instructions to prevent the agent from interpreting instructions contained within the text files it processes.
Capability inventory: The skill has the capability to perform network POST requests to the ElevenLabs API and write audio files to the local filesystem.
Sanitization: The script uses a slugify function to sanitize profile and voice names before using them in output filenames, which helps prevent basic path traversal attempts.

elevenlabs-tts