agents
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.90). The skill lets agents ingest and act on external, potentially untrusted content — e.g., arbitrary webhook URLs in tools (tools[].webhook.url), uploaded knowledge_base documents, custom LLM endpoints (prompt.custom_llm.url), and client/widget resources (external scripts on unpkg.com or avatar-image-url) — which the agent reads and uses in conversation.
MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).
- Potentially malicious external URL detected (high risk: 0.80). The embed snippet loads and executes remote JavaScript at runtime from https://unpkg.com/@elevenlabs/convai-widget-embed, which is a required client-side dependency for the widget and therefore runs remote code in users' browsers.
Audit Metadata