translation-novel

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow ingests outsider-authored free text when the user supplies/pastes the novel chapter .md (Layer 2 “Parse chapter” reads chapter .md as plain text, then Layer 2 “Scene chunk” and Layer 3 “Core loop” include the chapter scenes in the LLM prompts).