story-to-storyboard

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

• Third-party content exposure detected (high risk: 0.70). Runtime Step 0/Step 2 reads the user-selected narrative markdown files from the local filesystem (outsider category is any document the user did not author), then Step 2 onward feeds extracted narrative text (governing thought, section evidence) into the agent’s LLM context; no sanitization/injection hardening is described.