Paste a credit policy, covenant test or any rule table. The audit runs thousands of cases through it and scores the rules with the Oursler Automaton Equation: where no rule answers, where rules disagree, which rules never matter, and which decisions flip on a 1% change. Everything runs in your browser. Nothing is uploaded.
> >= < <= == !=, and, or, not and parentheses. A rule can use another rule by its id, as in and not R4. Order matters: the first rule that fires sets the decision, the way most policy engines work.| RULE | DECISION | FIRES | DECIDES | IN CONFLICT | STATUS |
|---|
The audit learns in three ways, with no code changes or redeploys. Auto-fix repairs gaps and conflicts one step at a time and keeps only the changes that raise the score. Learn from outcomes tunes each rule's thresholds against how past cases actually turned out, and keeps a change only if it also holds on cases it never trained on. Memory saves every run on this device so you can watch the rules improve and roll back to any version.
outcome column: either good/bad (repaid or defaulted) or the decision that should have been made.| WHEN | SOURCE | SCORE | CLARITY | COVERAGE | BRITTLE |
|---|