prompt-xray

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

• Third-party content exposure detected (medium risk: 0.65). Required workflow loads and parses extracted_results/*_extracted.json at runtime via xray_helper.load_prompts(), then uses the extracted JSON fields to generate Markdown that is fed into the agent/LLM context; if those JSON files contain outsider-authored prompt text, this becomes an indirect prompt-injection path.