The Agent Skills Directory

[SAFE]: The skill acts as a design reference for production voice applications and does not contain any executable malicious code or dangerous instructions.
[CREDENTIALS_UNSAFE]: The skill follows secure development practices by instructing users to manage secrets via environment variables (ELEVENLABS_API_KEY) and providing a configuration layer to load them securely, specifically warning against hardcoding credentials.
[EXTERNAL_DOWNLOADS]: The skill references official documentation and repositories from well-known services (ElevenLabs, sindresorhus, isaacs). These references are for legitimate dependency management and documentation purposes.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external text through the TTS service layer, which is the intended functional purpose.
Ingestion points: src/services/tts-service.ts via the generate and generateLongText methods which accept string input.
Boundary markers: None (typical for text-to-speech input).
Capability inventory: Network requests to the official ElevenLabs API (api.elevenlabs.io).
Sanitization: Includes logical splitting of text to manage API limits but does not filter content, as it is a specialized media generation task.

elevenlabs-reference-architecture