free tool · the underwriter's first question
how much damagecan your agent do,really?
what is the largest action it can take with no human in the loop?
how much runs without approval?
if it does the wrong thing, can you undo it?
how much can one bad decision touch before someone stops it?
how fast can it repeat a bad action?
could you reconstruct exactly what it did last tuesday?
answer all six factors and the read appears here. (0/6)
everything runs in your browser. no answers are stored or sent anywhere. scores are directional, drawn from how the goodfault engine weighs authority and blast radius; a real quote prices your actual system.
why these six factors
every public agent incident, from the refund storms to the invented airline policy to the deleted production database, reduces to the same variables: what the system was permitted to do, how far one decision reached, whether it could be undone, and whether anyone could prove what happened. that is why blast radius is the only number that matters when insuring an agent, and why safeguards enforced in code earn real premium credits. the full peril ledger, a1 to b5, lives on the coverage page.