The blue team wrote us a how-to-catch-us guide, thanks

lamia415 38 posts Since Jun 25, 2026

kaseya put out this MSP piece on catching AI phishing and honestly it reads like a list of our own mistakes. free after-action report

mort1cia 59 posts Since Jun 25, 2026

the recon section is just describing what we already do. scrape linkedin, feed the profiles to a model, get a personalized spearphish in minutes. they even cite a 54% click rate matching human experts

deephax 60 posts Since Jun 25, 2026

54%. at a fraction of the cost. the ROI on this is obscene

b1tfr34k 77 posts Since Jun 25, 2026

the part that matters is polymorphic. every email unique — subject, sender, formatting all rotating — so the signature-based gateways have nothing to match. that's why the filters miss

lilith 49 posts Since Jun 25, 2026

right but read the second half. they're not trying to stop it at the inbox anymore. they gave up on that. they're watching behavior after the click

mort1cia 59 posts Since Jun 25, 2026

yeah and that's the part we need to actually respect. the tells they list are all us being sloppy post-compromise:

mort1cia 59 posts Since Jun 25, 2026

dropping a mailbox forwarding rule to an external address right after login
impossible travel," same account two countries in minutes
hammering MFA push until they accept
logging in clean then immediately spraying powershell or touching systems that account never touches

deephax 60 posts Since Jun 25, 2026

every one of those is an own-goal. the phish is flawless and then people set an obvious forward rule from a russian IP thirty seconds later

lilith 49 posts Since Jun 25, 2026

the forwarding rule is what gets everyone. don't create noisy inbox rules. don't pivot the second you're in. sit. blend. match the user's normal hours

lamia415 38 posts Since Jun 25, 2026

impossible travel is why you route through infra in the target's region. logging in from their city instead of ours kills that alert dead

b1tfr34k 77 posts Since Jun 25, 2026

MFA fatigue works but it's loud — a pile of unrequested pushes is literally on their list now. one clean prompt-bomb attempt, not forty

deephax 60 posts Since Jun 25, 2026

the meta lesson: their detection isn't about the email anymore, it's correlation across identity + endpoint + email. so our job is to not light up more than one of those at once

lilith 49 posts Since Jun 25, 2026

and speed. they're measuring response time now, dwell time is the whole game for them. which means it's the whole game for us too — get in, get what we came for, don't linger admiring the view

mort1cia 59 posts Since Jun 25, 2026

"one eventually gets past the gateway." they said it themselves. the inbox was never the wall. the wall is whether anyone's watching what happens next

lamia415 38 posts Since Jun 25, 2026

so watch what happens next. quietly.