Your adversary,on your side.
Someone is already trying to hack you — your inbox, your money, the people who trust you. Feint is a team of people and AI agents who hack you first, with your written consent. You find the open door in a rehearsal, not in a loss you can't undo.
“The target was never me. It was my assistant's inbox.”
The question
When the real attempt comes, will it find you ready?
Everyone now has a motivated, tireless, automated adversary. Almost no one has a defender who works the same way. We are the attempt you asked for — run quietly, on a schedule, by people who like you. So the one you didn't ask for finds you already prepared.
Every night, while you sleep
One night of the fleet
A fleet of AI agents maps your world, hunts for the open door, and then hands every candidate to a second agent whose only job is to disprove it. The one who checks a finding is never the one who found it. By six in the morning, a short report is waiting.
The counts below come from one of our own deep audit runs: 55 things to check, 37 that survived to the final record. Most candidates are rejected on purpose.
The full methodThe part a machine can't do
When it matters, a person calls you.
The fleet works all night. It never tires, and it never picks up the phone. When a finding is real, someone who knows your household calls you in the morning — a voice, not a ticket, not an alert in an app you'll never open.
Underneath the machines, this is a human business. People test people. People make the call, and send the note, and sit with you while you decide what to do. The trust is the product, and trust is built one conversation at a time.
Meet the people on your sideWhat you get
Agents for breadth. People for judgment.
The fleet watches everything, every night, cheaply. People do what agents can't: chain the small things together, test a person's habits with care, and decide what actually matters.
The nightly fleet
Your devices, accounts, and home network, audited every night on private models. A plain-language report each morning.
Every tierExposure monitoring
What an attacker could learn about you before they try. Breach alerts, data-broker removal, a quarterly dossier on yourself.
Parry +Human penetration testing
People test what agents miss. Your home network, your recovery flows, your travel devices, on a schedule.
Riposte +Consented bouts
We rehearse the moment that matters with your household, so a real attempt meets a practiced response. Always with a stop word.
Riposte +A dedicated cell
For one family: a team of five people and the fleet, working year-round inside a signed consent envelope.
By introduction
A story · illustrative composite
“They practiced the call with me in the spring. When the real one came, I knew to ask about the olive tree. He hung up.”Ruth Okafor-Bell, 78 — Riposte member
AI voice-cloning turned the old ‘grandchild in trouble’ scam into something that sounds exactly like family. A rehearsed question is what holds. We practice it with you.
Read the storiesThe journal
What we're watching.
Every attack we teach against is real and already happening. We write each one down in plain language, so you can act on it before it reaches you.
Attackers hired agents. So should you.
AI now runs most of a real intrusion, and automated phishing is clicked four times more often.
Oct 2026The quiet collapse of the basics.
Regular use of two-factor fell from 94% to 53% in three years. The gap is follow-through.
Sep 2026You built it in a weekend. Who checked it?
Scans keep finding thousands of AI-built apps leaking real customer data. The right questions to ask.
Aug 2026The call that sounds like family.
A cloned voice needs only seconds of audio. One old habit, a shared question, defeats most of it.
Jul 2026