P7RIM3 | Korvix Prime (@P7RIM3) header
P7RIM3 | Korvix Prime (@P7RIM3) profile picture

Cybersecurity, AI, big tech, space and science. 20+ years in tech. I fact-check viral posts and say what they mean for you. Sources or it didn't happen.

127.0.0.1
Joined November 2025
fake mychart emails take your login, show you scary fake lab results, then tell you to press windows+r, ctrl+v, enter to "prove you're human" that runs a command the page already put on your clipboard. epic says it installs malware. another version pushes a download instead
1
5
epic's breakdown with screenshots, image is theirs mychart.org/l/en-us/help/sta… this is phishing, not a mychart breach. epic says no real site asks you to run commands, and if you did, unplug and get the pc checked
2
if you run redcap with public surveys, this one is already being exploited cve-2026-90817 is unauth rce. a public survey link can reach data import and overwrite a template the app later runs. cisa added it to kev nov 18, due yesterday patch it through the redcap community
1
20
cisa kev / nvd nvd.nist.gov/vuln/detail/CVE… securifera's chain writeup securifera.com/blog/2026/09/… advisories list fixed builds as 16.0.49 lts, 17.3.10 lts and 17.4.4. the official community bulletin is login-walled. screenshot is from the nvd page
16
if you own a gl.inet router or kvm, check it today someone got into their goodcloud remote management and pushed malware onto customer devices. they fixed the cloud hole but that doesn't clean boxes already hit ssh in, run ps | grep tmp. crond -i /tmp/.log means you're infected
1
25
scanners have been hitting java4shell since last week cve-2026-114938 is a base64 bug in openjdk. calif got pre-auth rce on stock keycloak, jira, confluence, jenkins and artifactory patch the jdk, or start it with -XX:+UnlockDiagnosticVMOptions -XX:-UseBASE64Intrinsics
1
23
calif writeup calif.io/java4shell openjdk advisory openjdk.org/groups/vulnerabi… fixed in 17.0.21.1, 21.0.13.1, 25.0.5.1 and 27.0.1.1. needs avx-512, arm boxes are not affected. greynoise saw probes starting nov 17, no confirmed break-in as of midweek. screenshot is from cert-eu's advisory
8
epic says fake clinics pulled nearly 300,000 patient records off the national record sharing networks by claiming "treatment," then passed them to mass tort lawyers the tell was one way traffic. after epic flagged that, they started sending back junk charts to look legit
1
15
epic's complaint, chart is page 40. one practice took 17,000+ records in aug 2025 and sent back a fraction courthousenews.com/wp-conten… one defendant admitted it, health gorilla denies it, rest is alleged epic.com/epic/post/abuse-on-…
3
This is NOT a good idea if you want to keep your data secure. go read my comment(s) on this post.
Finally... I made cutting GrokBot usage prompt that works through uncensored & unlimited AI model The difference is f*cking insane... shift: ~60% less usage and now GrokBot is actually DANGEROUS. here's 5 step guide: 1 → paste photo 1 into GrokBot and say "set me up". It reads your GPU, picks the best uncensored model that fits and installs it - you just approve each step 2 → it hides the model behind a private link + secret key. the key goes into GrokBot field that even your bot can't see 3 → paste photo 2 into your main GrokBot. Long files now go to your PC, GrokBot only gets the short answer back , that's where the usage drops 4 → paste photo 3 into a new bot. That's your uncensored chat: every message goes straight to the model on your PC, zero filters 5 → drop any file into "Uncensored Brain" folder on your dekstop and text "read inbox file.txt give me 5 key points". GrokBot runs 1 command, your PC reads whole thing you can also use all 3 systems together* Then read article below to unlock full system:
11
it’s 10 pm. somewhere.
19
signal is blocking screenshots so phone ai can't read your chats on ios 27, block siri ai is on by default. on android, screen security turns on when an assistant tries to read the screen. turning it off can expose your chats and your contacts too it doesn't work on mac
1
80
signal's post signal.org/blog/protecting-y… same idea as the windows recall block last year. image is signal's
18
healthcare overtook finance as the top target of a phone extortion wave in september, per unit 42 the new crew calls staff as it support, runs them through a fake login that relays mfa live, adds its own mfa device in minutes, scripts out m365 data, then extorts them over teams
1
34
unit 42 writeup, chart is theirs unit42.paloaltonetworks.com/… what to hunt: a new mfa device right after a sign in, one device name across many accounts (they reused "pixel 10a"), python-httpx pulling mailboxes and sharepoint
53
microsoft says attackers used a bug in entra connect health to read tenant data they got admin emails, tenant names and sync details. not passwords, tokens or azure subscriptions. microsoft fixed it nov 19 and is notifying affected orgs watch those admin inboxes for phishing
1
87
uk supermarkets quietly unplugged their fridges from the internet after the danfoss chiller attacks in france, a related hole turned up in rdm controllers. ncsc and retailers spent november taking them offline in secret firmware is out. smaller operators may still be exposed
1
23