kelly-drama

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, the agent ingests user-/outsider-authored free text from Busabase “shots” fields (notably prompt, negative_prompt, action, composition, camera, setting, lighting, and srt) when building the storyboard prompt in app/app/js/drama-model.js (storyboardPrompt/storyboardPromptPreview) and then sends it to external image/video/voice generation via the provider request flows in app/app/js/actions.js/providers.*.