Guides.
Evergreen explainers on outflow control, AI agent governance, SOX/ICFR refund controls, and shadow mode. Written for the buyer figuring the category out for the first time.
Read by topic.
A Morse code message moved $174,000. What would have stopped it, and what would not.
In May 2026 an attacker posted a Morse code message on X. An AI agent decoded it, a second agent treated the decoded text as an authenticated instruction, and roughly $174,000 in tokens left a verified wallet. No contract bug, no stolen key. A walkthrough of the failure, and an honest account of which controls would have caught it.
Read guide →Guide · 6 minMake the gate real: wiring meridian-company-os to a production spend gate
We took the spend gate from meridian-company-os, the repo behind the viral Company OS guide, and wired it to a production decision engine. Same inbox, same UX, one change: a spend approval only lands after an external policy engine says yes.
Read guide →Guide · 6 minGoverning agent spend while the GENIUS rules keep moving
The GENIUS Act's statutory rulemaking deadline passed on July 18, 2026 with no final package, and the open comment periods run into late August. How to govern agent and stablecoin spend now, so that whatever text finally lands, you can show which policy version governed each transaction.
Read guide →Guide · 6 minx402 pre-authorization, explained: signed tokens before agent money moves
x402 removes the human from the payment loop by design. Pre-authorization puts a policy decision back in: the agent's payment becomes an intent, the engine evaluates it, and a signed, single-use, scope-bound token is the only thing that lets money move. No token, no payment.
Read guide →Guide · 5 minWhat outflow control is, and why refund control was the wrong noun
Outflow control is the policy and approval layer that sits in front of every outbound money movement on Stripe: refunds, credits, payouts, transfers, dispute responses. Why the broader noun matters now that AI agents can move money.
Read guide →Guide · 6 minYour AI support agent is now a financial actor. Govern it accordingly.
Once an AI agent has refund permissions, the policy layer that used to sit in a CS supervisor's head needs to live somewhere the agent can be evaluated against. Practical guide to wrapping autonomous refund decisions in policy.
Read guide →Guide · 7 minRefund controls under SOX 404 and ICFR: the smallest defensible setup
What auditors actually look for when refund and payout surfaces show up in a SOX 404 walkthrough, and the minimum control set that holds up. Pragmatic, not aspirational.
Read guide →Guide · 5 minShadow mode: what you actually learn from 90 days of replay
Connecting Stripe read-only and replaying your last 90 days of activity against a proposed policy is the cheapest way to find out what your refund policy is actually doing. Walkthrough of what the report tells you.
Read guide →Want to see outflow control on your own Stripe data?
Connect Stripe read-only and replay your last 90 days against a draft policy. Shadow mode is free.
Start in shadow mode first. Move to live enforcement later.