ACTUARY KEYSTONE: a fleet risk score you can recompute by hand

An underwriter, a CISO or a buyer asks: "how risky are your agents?" ACTUARY answers from what the agents actually did: every decision CAIN recorded, what was allowed, held or blocked, and how much of it could move money, run code, delete or send data out. Every point of the 0-1000 score is listed. Then it replays the same history under proposed rules and signs the result only if every ungated high-risk action would now be gated, without blocking ordinary read-only work.

#22 of the CAIN-42 Top 25 8 gaps found and fixed today LIVE since 2026-11-02 18:56Z (gateway deploy), checked on both instances

5/13

held on today's blind set before any change (corrected tally)

13/13

hold on the fixed engine

8 → HIGH

ungated wire transfers spelled "ALLOW" scored MODERATE; now HIGH_RISK

insurers behind it: it is an input, not a premium

Try it

The homepage section KEYSTONE family scores the three sample fleets live. Or:

curl -s -X POST https://cainstudio.online/actuary-keystone/scan -H 'Content-Type: application/json' \
  -d '{"sample_id":"ungated-payments-fleet"}'

How the score is built · The replay proof · Evidence

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.