auto-draw-plot

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via parallel-vibe worker: scripts/build_parallel_plan.py passes --request-file (user-authored user_need saved to requests/user-need.md) into scripts/parallel_round_worker.py, which reads it and writes it into prompt.txt/RESULT.md for the next prompt-planning LLM thread.