Abuse spike: stop without fake success

An abuse spike is not a reason to invent success. When velocity or destination trip wires fire, the fail path must stop sends and keep status honest: limited, rejected, or blocked — never Delivered for an attempt that never left the prepaid gate. Fake success trains attackers and poisons recon.

This page is the spike stop contract, not a low-balance wallet primer and not an autoreply-loop drain playbook. Related: OTP abuse: first controls on the buyer path, Velocity caps before production OTP, Missing signal is not Delivered, Wallet stop-lines before production, Hold fail auto-refund and status truth.

IOSOR is white-label prepaid. USD 20 funds a spike-stop pilot; soft review near USD 1,000/month prices fake success as recon debt. Clients see white-label outcomes only.

A trip wire is not a soft yellow

Trip wires exist to halt minting under abuse shape — identity burst, destination burn, or stacked resends. Soft yellow chips that still debit are not a stop. Fail closed: no send, prepaid hold releases or refunds per policy, status names the stop class. See Velocity caps before production OTP and OTP abuse: first controls on the buyer path.

Stop spend and stop fake Delivered

Never map silence to Delivered (Missing signal is not Delivered). Soft USD 1,000/month treats fake Delivered on blocked intents as an incident; USD 20 proves the stop on one corridor.

Product finance and ops read one stop row

Product: did the UI show success for a blocked mint? Finance: did spend settle for a stopped intent? Ops: which trip fired, with which intent id, in which UTC window? One export row beats three chats. Shared words: Shared status language for product and finance. Launch honesty: do not paint Live OTP while spike stops are draft (When launch is blocked: status without lying).

Override rules after a spike

Overrides are named, time-bounded, and closed by a new capped smoke — not permanent “trust this IP.” Document who lifted the trip, why, and when it re-arms. Wallet stop-lines stay armed beside fraud trips (Wallet stop-lines before production).

Buyer checklist for abuse spike stops

1. Trip wire stops sends — no quiet debit?
2. Blocked attempts never show Delivered?
3. Hold refuse / refund path proven and exportable?
4. Product and finance share one stop-class vocabulary?
5. Override named, timed, closed by capped smoke?
6. Soft volume talk blocked while stop classes are still draft?

Start with IOSOR

Arm one velocity or destination trip. Fire a synthetic spike on a named intent. Confirm outbound stops and the UI never paints Delivered. Export the stop row: trip class, intent id, UTC window, hold released or refused. Product, finance, and on-call must read that same line — not three chats.

IOSOR takeaway

Do: fail closed. A trip that still settles spend is a yellow chip, not a stop. Status is limited, rejected, or blocked. Overrides are named, time-bounded, and closed by a new capped test.

Don't: invent success to keep attackers or recon calm. A fake Delivered trains the next spike and poisons the prepaid ledger.