animated-explainer-ad

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The runtime workflow uses the operator-provided free-form kickoff prompt (Phase 1 “Intake” → extract brand/product/problem/mechanism/audience → draft project/brief.md and scene VO text), so that outsider text is ingested into the agent/LLM context and then used to generate the ad (direct indirect injection surface via user text, not from public web content).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime scripts call remote model APIs (FAL and ElevenLabs) that generate images/audio/video used directly in the pipeline (e.g., https://api.elevenlabs.io/v1/music and the ElevenLabs TTS endpoint https://api.elevenlabs.io/v1/text-to-speech/{args.voice_id}), so external services execute remote model code that directly controls the agent's outputs.