skills/guyronnen/openmontage/agents/Gen Agent Trust Hub

agents

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to install official ElevenLabs packages and CLI tools from NPM and PyPI, including @elevenlabs/cli, @elevenlabs/elevenlabs-js, and elevenlabs.
  • [INDIRECT_PROMPT_INJECTION]: Agents created using these instructions ingest untrusted user voice and text input. The skill documents a broad capability surface including webhook calls, browser-side JavaScript execution, and outbound telephony.
  • Ingestion points: User audio transcripts and text messages processed by the conversational AI agent.
  • Boundary markers: System prompts defined in conversation_config and built-in platform guardrails mentioned in agent-configuration.md.
  • Capability inventory: Webhook execution (POST/GET), client-side tool execution in the browser, and outbound Twilio phone calls.
  • Sanitization: Built-in content moderation guardrails and the sanitize option for dynamic variables in tool responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 12:09 AM
Security Audit — agent-trust-hub — agents