Sergi Ortega (@x0rPE) header
Sergi Ortega (@x0rPE) profile picture

Threat Intelligence, Threat Hunting, Incident Response and RE/Malware Analyst, that's me.

between call and retn
Joined July 2012
Photo by Sergi Ortega (@x0rPE)Photo by Sergi Ortega (@x0rPE)Photo by Sergi Ortega (@x0rPE)Photo by Sergi Ortega (@x0rPE)Photo by Sergi Ortega (@x0rPE)Photo by Sergi Ortega (@x0rPE)
Sergi Ortega retweeted
We have identified interesting features of #Aigarth's #AI training algo: 1. It creates hypotheses, verifies them, drops wrong ones, returns to put aside ones; 2. It uses the same pattern for different phenomena if they have similarities, applies little modifications to the pattern if necessary; 3. It chains simple calculations into complex ones, rearranges them. All this looks similar to how hoomans reason.
32
206
18
746
42,927
Sergi Ortega retweeted
$50 #QUBIC GIVEAWAY — #QUBIC HOLDERS ONLY. To enter: 1) Follow @Adi_bulls Rt and like❤️ 2)Drop your #Qubic address No multi-bag hoppers. True $Qubic wallet only Winner gets $50 in $QUBIC. Feeless. Instant. Diamond hands only. winner in 24 hours 💸⏳️⌛️⏰️
🤖 Made with AI
22
24
2
59
1,441
Sergi Ortega retweeted
#AI Shut Down Exposes #Centralisation. 
#OpenAI, #Anthropic, #Google, #xAI, and #Meta each run their #AI in huge, #centralized data centers. 🏢🏢 #Nvidia sells most of the chips those factories use. When #ChatGPT, #Claude, #Grok, and #Gemini all went down at the same time, it showed the weakness.📵 A lot of the world’s #AI still lives in a small number of buildings, clouds, and networks. Why that can happen to them… • They all rent space from the same few cloud giants (think #Amazon, #Microsoft #Azure, #Google Cloud). If one of those clouds or a major internet pipe has a problem, several #AIs can wobble together. • Their models live in massive clusters. One bad config, one overloaded region, one network glitch, and the whole product goes “sorry, try again.”📴 • One company; one control room. A software update, a security lock down, a billing fight, or a single team’s mistake can take the service offline for millions of people at once. • They compete, but they still share the same plumbing… Chips, power, fiber, CDNs. That’s why “different logos, same outage window” is possible. That’s not a conspiracy. It’s just how #centralized systems work. One factory goes dark, the product goes dark. #Qubic is built the opposite way. #Qubic is not another chatbot company sitting in one data center. It is a public network where thousands of independent computers around the world do useful work. 🖥️🌏 That work trains #AI (their project is called #Aigarth) instead of wasting energy on empty puzzle solving. The network is run by 676 computers called Computors. Nothing important is accepted unless at least 451 of them agree. No single CEO, no single cloud account, no single office can flip the switch off. Why “everyone’s #AI went down at once” doesn’t map onto #Qubic the same way… • There is no one #ChatGPT style app owned by one lab that the whole world is logged into. The compute is spread across many machines owned by many people. • If one miner or one Computor dies, the others keep going. That’s the whole point of a distributed network. • The #AI training is the mining. ⛏️ You don’t have “the #AI server farm in Virginia” as a single point of failure the way #OpenAI or #Anthropic do. • No company can “walk away from a partner,” change the terms overnight, or take the model offline because of a board fight. The rules live in the protocol. Could #Qubic ever have problems? Yes. Software bugs, not enough participants, or a bad upgrade can still hurt any network. What it doesn’t have is “the one company that runs the world’s favorite #AI just went down and took half the internet’s brain with it.” The last 24 hours were a live demo. The biggest #AI brands can all stumble together because they are still #centralized products sitting on shared pipes. #Qubic’s bet is different… Turn the world’s leftover computing power into a public #AI engine that no single company owns. Miners get paid. The work actually trains intelligence. Tokens get burned when the network is used. 🔥 And the system is designed so one outage at #Microsoft or #OpenAI doesn’t turn the lights off for everyone. Big Tech is racing to build smarter gods in private boxes. #Qubic is trying to grow intelligence across a public network. When the private boxes all blink at the same time, the idea of “#AI that isn’t owned by one company” suddenly looks a lot less like a #crypto slogan and a lot more like a backup plan. #DYOR Study @ Qubic.org #Qubic #Decentralised #Intelligence #Ai
2
19
1
84
1,471
Sergi Ortega retweeted
Scales, the eBPF malware targeting ArchLinux sha0coder.github.io/scales/
1
16
1
77
8,671
Sergi Ortega retweeted
New HackTricks tools at tools.hacktricks.wiki/: - Request to update outdated HT pages with the researcher tool - Use the API for RAG-ready best matches from hacktricks - Access the HT AI chatbot via API - ... The new Pentester plan is live, and every tool includes a free tier.
1
3
20
1,383
⚠️ A new ClickFix variant abuses Win+R to mount a remote WebDAV drive and run malware. It launches a trojanized WorkFlowy Electron app that beacons to C2 every 2 seconds. @Atos says it bypassed Microsoft Defender and surfaced only through threat hunting. 🔗 Inside: WebDAV trick + ASAR injection → thehackernews.com/2026/03/in…
10
64
4
237
29,507
Google & partners disrupted IPIDEA, one of the world's largest residential proxy networks, reducing its device pool by millions. This infrastructure was leveraged by over 550 distinct espionage and cybercrime groups. Full report + IOCs here: bit.ly/4a7hT9h
21
46
6,062
🦖An interesting Velociraptor artifact to scope Virtualisation Worker process and looking at handles to disk images. We are interested in unusual location or small disk size. 🔎Artifact: docs.velociraptor.app/exchan… 🛡️Reference: bitdefender.com/en-us/blog/b… @velocidex
1
15
606
Sergi Ortega retweeted
My team colleague, Yann Malherbe, wrote an in-depth blog post about the Automation of VHDX Investigations with 🦖. Enjoy. labs.infoguard.ch/posts/auto…
11
35
3,088
Sergi Ortega retweeted
🚨 New Report: #TA BQTlock #Ransomware Analysis Our #Threat Research team breaks down this emerging hacktivist-criminal #RaaS (+90 pages) 👇👇👇 🔗Report: socradar.io/resources/whitep… #malware #CTI #reverseengineering #RE #MalwareAnalysis
5
30
3,670
Sergi Ortega retweeted
I just released MFTool, an NTFS parser that builds an in-memory map of a volume, allowing you to: - Read any file without opening a handle - Get the contents of locked/deleted files (registry hives, pagefile.sys, etc) - Perform fast, in-memory searches across the entire disk 🔗👇
4
43
164
13,728
Sergi Ortega retweeted
Después de mucho tiempo.. Nueva herramienta y entrada en el #blog!🙃 NekoDNS: Jugando con DNS una vez más: darkbyte.net/nekodns-jugando… NekoDNS - Experimental Reverse DNS Shell: github.com/JoelGMSec/NekoDNS #Darkbyte #Hacking #DNS #ReverseShell #NekoDNS
2
16
55
5,054
Interesting report from Sophos covering malicious Velociraptor use 🔎 news.sophos.com/en-us/2025/0… Checking out this msi: 💾 • they used Velociraptor version 0.73.4. • Server likley installed on ~04/08/2025 10:03:15 (self signed certificate) • v2.msi - virustotal.com/gui/file/649b… As the article suggested unexpected processes communicating to workers\.dev would detect this particular instance. I also created a Velociraptor inception artifact a while ago to find unauthorised instances that uses yara and other methods github.com/rapid7/Rapid7-Lab… @velocidex
21
49
3,959
Sergi Ortega retweeted
RIP to all the detection rules leveraging RunMRU. Time to go back to the drawing board!
A Windows #Clickfix alternative seen in the wild on a mass-spreading malware campaign bypassing traditional Win+R shortcut restrictions User is asked to open the Windows Power User menu (Win+X), open a Powershell terminal and paste and running a malicious Clickfix-style command
2
3
65
5,124
Sergi Ortega retweeted
𝗭𝗲𝗿𝗼 𝗖𝗹𝗶𝗰𝗸, 𝗢𝗻𝗲 𝗡𝗧𝗟𝗠: 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗣𝗮𝘁𝗰𝗵 𝗕𝘆𝗽𝗮𝘀𝘀 A newly discovered zero-click vulnerability, CVE-2025-50154, bypasses a Microsoft patch, allowing attackers to steal NTLM hashes without user interaction. Microsoft has issued patch in August Patch Tuesday and here's a KQL to detect this bypass. .🫡 cymulate.com/blog/zero-click… msrc.microsoft.com/update-gu… Detection: detections.ai/share/rule/Crr… #cybersecurity #vulnerability #NTLMLeak
7
85
2
293
39,815
Sergi Ortega retweeted
#APT #Patchwork #DroppingElephant #WhiteElephant #threat #malware #BADNEWS 📍🇮🇳 💥🇵🇰🌏 ⛓️ #Phishing > LNK | XLSX (MSC) + JS > Download > Bait doc + exe > Task > exe extract dll + Load > #RAT > Info Collection > #C2 (Mythic) 🔗360 Threat Intelligence: mp.weixin.qq.com/s?__biz=MzU…
1
18
81
10,796
Sergi Ortega retweeted
I didn't want to drop a new *Fix naming variant on you guys, but there you go. All jokes aside, it was a great collab with @Octoberfest73. Awesome person and really helps you push the limits when researching and testing💪
This quoted post is unavailable.
2
24
90
10,112
Sergi Ortega retweeted
🚨#DarkCloud, New research of this #malware #stealer. Seen in numerous campaigns, hits companies and users for all kinds of information🚨 🔗Full Analysis: rexorvc0.com/ #RE #intel #threat #malwareanalysis #infosec #cti
5
13
1
38
12,506
Sergi Ortega retweeted
New samples of previously-unseen UMBRELLA STAND & SHOE RACK malware just landed on VirusTotal! These custom implants target Fortinet FortiGate firewalls, reinforcing a growing trend we've seen during the past years: router exploitation as a long-term foothold and stealth access. UMBRELLA STAND: * Hooks reboots, uses ldpreload for stealth * AES-encrypted fake TLS C2 * Attackers deploy BusyBox, tcpdump, nbtscan... SHOE RACK: * Reverse SSH tunnel over fake SSH-1.1.3 * Uses DNS-over-HTTPS to locate C2 * Based on modified open-source Go tooling Both tools are typically found together, likely used by an unknown China-nexus actor (also see the famous Dutch COATHANGER report and overlap). BTW, great job by @NCSC for sharing the detailed teardown, IOCs, YARA rules and now the samples!
Sample is now on VT! 🚩Hash: ff82ea16717c6fa9fc8c07ae2de09c5f 🎯Actor name: UMBRELLA STAND 🔹Comment: UMBRELLA STAND has been observed targeting internet-facing, FortiGate 100D series firewalls produced by Fortinet. Several of the tools described in the COATHANGER report were also present on this device. A previously unknown piece of malware… 🌐URL: ncsc.gov.uk/section/keep-up-… 🔎OnVT: virustotal.com/gui/file/ff82…
1
35
1
95
16,093