website-to-hyperframes

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

• Third-party content exposure detected (high risk: 0.78). Outsider-authored free text can enter the LLM context via Step 0’s runtime website capture: npx hyperframes capture <URL> downloads pages and extracts readable text (e.g., capture/extracted/visible-text.txt / asset-descriptions.md) from the provided public URL, which the workflow later reads in Step 2/3 to draft the script/storyboard.