Verify it yourself

1. Download the data:

   curl -sO https://cainstudio.online/falsework-evolution-2026-11-02.json

2. Send every blind case to the live scanner and compare grades (stay under 30 per hour):

   python3 - <<'EOF'
   import json, time, urllib.request
   d = json.load(open("falsework-evolution-2026-11-02.json"))
   for c in d["blind_c"]["cases"]:
       req = urllib.request.Request("https://cainstudio.online/keystone/scan-source",
           data=json.dumps({"name": c["file"], "source": c["source"]}).encode(), headers={"Content-Type": "application/json"})
       got = json.load(urllib.request.urlopen(req))["grade"]
       print(c["kind"], c["name"], "published:", c["grade_after"], "live:", got)
       time.sleep(3)
   EOF

   Until the next deploy the live engine is the older one, so a few grades will differ: that difference is the change.
3. Send each case twice. The answer must be identical (the engine is deterministic).
4. Write your own fakes. The most useful thing you can tell us is a fake it grades REAL. We will publish it.

FIRME AI CORPORATION · CAIN-42 is a Byzantine trust fabric. Numbers marked "our own check" were run by us, not by an independent party. Nothing on these pages is a certification.