RedHub AI
How the roll-up gate works
How the Validation-Layer Coverage Map works
This tool grades every automated decision across four validation layers, discounts any layer too thin to count as a layer, then rolls the map up to its weakest decision. It reads 8 decisions and rolls them into one verdict — the worst, never the average — then a gate that can only make it worse.
The 8 decisions
A layer that covers fewer than 25% of decisions is a pilot, not a layer — it is marked NOMINAL and stops counting toward any decision's depth, so a decision can have all four cells full and still not read FOUR-LAYER. The map then takes its worst decision: one unvalidated decision reads BLIND SPOTS however high the fill rate.
Why: The map is 63% filled — on the fill rate alone it looks two-thirds done. But reading down the columns, the output-checking layer reaches only 13% of decisions: below the 25% threshold it is NOMINAL and stops counting. That is why D-01 has all four cells full and still reads THIN — depth 3, not 4. Six decisions read THIN and two read UNVALIDATED, and one unvalidated decision reads BLIND SPOTS regardless of fill. No decision can reach FOUR-LAYER until output-checking crosses the threshold.
No single decision is the crisis — the clustering is. The roll-up takes the worst, the gate escalates when trouble compounds, and it names the one thing to fix first. Grades the map you draw, never people. A working aid, not legal, security, or compliance advice, and not an audit — MAP COVERED means what you drew came back covered, not that you drew everything.