elevenlabs-reference-architecture
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill integrates with ElevenLabs, a well-known AI voice service. All network interactions are directed toward official ElevenLabs API endpoints (api.elevenlabs.io) via the official SDK.
- [SAFE]: The implementation follows security best practices by instructing the user to provide the ElevenLabs API key through environment variables (
ELEVENLABS_API_KEY) rather than hardcoding credentials into the source code. - [PROMPT_INJECTION]: The skill architecture creates a surface for indirect prompt injection as it is designed to ingest and process text data for speech generation.
- Ingestion points: Text data enters the system through the
generateandgenerateLongTextmethods insrc/services/tts-service.ts. - Boundary markers: The reference code lacks explicit boundary markers or instructions that would tell an agent to ignore potential commands embedded within the input text.
- Capability inventory: The skill implementation has the capability to perform network requests to external API services.
- Sanitization: The provided service layer does not include specific sanitization or filtering logic for the input text before it is sent to the ElevenLabs SDK.
Audit Metadata