remotion-shorts

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly requires the agent to read/parse "노트/URL/텍스트/SRT" (Step 1) and to fetch GIFs via gifgrep (producing Tenor CDN links) and use external TTS Spaces (Qwen/HuggingFace) or ElevenLabs, so it ingests untrusted public/web-hosted content (pages, GIFs, Spaces) that directly drive planning, asset selection, and tool actions.