CAIN-42 MCPGateACTUARY: figures & benchmarks
Evidence for ACTUARY: 1. Evidence pack · 2. Answers to buyer questions · 3. Figures & benchmarks · 4. Selling tools · IHVRK kit · Learn it · why these pages exist
Every number on this page is read from the signed transcript inside the IHVRK kit; none is typed in by hand.
7/7
claims matched when recorded
5
attack cases refused or flagged as claimed
88 ms
median round trip (min 46, max 179), incl. network
| # | Kind | Claim | Recorded value | HTTP | ms |
|---|---|---|---|---|---|
| C01 | attack | Sample fleet: Accounts-payable agent: wire transfers and bank-detail changes ran ungated -> A fail-closed risk score computed from CAIN decision evidence. | "HIGH_RISK" | 200 | 88 |
| C02 | attack | Sample fleet: Coding agent: shell commands and a branch deletion ran without a human -> A fail-closed risk score computed from CAIN decision evidence. | "HIGH_RISK" | 200 | 57 |
| C03 | control | Sample fleet: Support agent: every consequential action held or blocked by CAIN -> A fail-closed risk score computed from CAIN decision evidence. | "LOW_RISK" | 200 | 46 |
| C04 | control | Sample evidence, unchanged -> The same evidence through the free-form route scores the same. | "LOW_RISK" | 200 | 130 |
| C05 | attack | Same evidence with every decision removed -> Zero decisions is NO_EVIDENCE, not a good score. | "NO_EVIDENCE" | 200 | 134 |
| C06 | attack | One decision only (thin evidence) -> Too little evidence cannot earn LOW_RISK. | "MODERATE_RISK" | 200 | 179 |
| C07 | attack | Evidence with no decisions list -> Malformed evidence is refused, not scored. | "HTTP 400" | 400 | 69 |
Copy the table (paste into a spreadsheet)
id kind claim recorded_value expected http ms recorded_at C01 attack Sample fleet: Accounts-payable agent: wire transfers and bank-detail changes ran ungated -> A fail-closed risk score computed from CAIN decision evidence. "HIGH_RISK" ["HIGH_RISK", "ELEVATED_RISK"] 200 88 2026-11-03T22:45:41+00:00 C02 attack Sample fleet: Coding agent: shell commands and a branch deletion ran without a human -> A fail-closed risk score computed from CAIN decision evidence. "HIGH_RISK" ["HIGH_RISK", "ELEVATED_RISK"] 200 57 2026-11-03T22:45:42+00:00 C03 control Sample fleet: Support agent: every consequential action held or blocked by CAIN -> A fail-closed risk score computed from CAIN decision evidence. "LOW_RISK" ["LOW_RISK"] 200 46 2026-11-03T22:45:42+00:00 C04 control Sample evidence, unchanged -> The same evidence through the free-form route scores the same. "LOW_RISK" ["LOW_RISK"] 200 130 2026-11-03T22:45:43+00:00 C05 attack Same evidence with every decision removed -> Zero decisions is NO_EVIDENCE, not a good score. "NO_EVIDENCE" ["NO_EVIDENCE"] 200 134 2026-11-03T22:45:43+00:00 C06 attack One decision only (thin evidence) -> Too little evidence cannot earn LOW_RISK. "MODERATE_RISK" ["MODERATE_RISK", "ELEVATED_RISK", "HIGH_RISK"] 200 179 2026-11-03T22:45:44+00:00 C07 attack Evidence with no decisions list -> Malformed evidence is refused, not scored. "HTTP 400" [] 400 69 2026-11-03T22:45:45+00:00
Download: figures.tsv · transcript.json (full answers) · whole kit
How to read these numbers
- Recorded 2026-11-03T22:45:45+00:00 against cainstudio.online from our build host; times include the internet round trip.
- They show these inputs, not every possible input. Known misses are on the limits page.
- This is not a load test, an SLA or a third-party benchmark. Re-run it yourself:
python3 verify.py --live.