Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj) header
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj) profile picture

Linux, Hacking, Security

Thailand
Joined March 2009
Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)Photo by Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens (@materaj)
Summary step by step for HELK setup and send log from Windows + Linux. github.com/materaj2/HELK_Set…
1
4
17
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
Unauth remote → kernel panic in ~5s on any open TCP port. NoisyTCP / CVE-2026-99288: 11-year race in Linux TCP established ehash lookups. • Stale ehash + request-socket free/realloc across RX queues (IPv4 frags vs normal TCP) • __inet_lookup_established returns unhashed socket → NULL listener deref in tcp_v4_rcv • Hits default Nginx / Apache / SSH. Container crash takes the whole host • Fixed in 6.12.117 / 6.18.63 / 7.3.4. Most distros still shipping vulnerable kernels Writeup + details: github.com/V4bel/NoisyTCP oss-sec: seclists.org/oss-sec/2026/q4… Credit: Hyunwoo Kim (@v4bel) sysctl net.ipv4.tcp_syncookies=0 is a temporary mit. Patch + reboot. #Linux #Kernel #DoS #CVE
2
5
33
2,029
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
🚨 A Next.js middleware flaw can let unauthenticated attackers bypass authorization checks; CVE-2025-29927 PoC: github.com/alinaclark6723772… The vulnerability affects applications that rely on Next.js middleware for access control. By abusing the x-middleware-subrequest header, an attacker can cause middleware checks to be skipped and potentially reach protected routes without authentication. Next.js Middleware | Authorization Bypass: • CVSS: 9.1 Critical • Attack: Remote, unauthenticated • Affected: Next.js 11.1.4–12.3.4, 13.x < 13.5.9, 14.x < 14.2.25, 15.x < 15.2.3 • Fixed: 12.3.5, 13.5.9, 14.2.25, 15.2.3 • Impact: Bypass of middleware-based authorization controls The repository demonstrates the header-based bypass against a protected route.
17
129
15
850
77,824
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
Sharing ARTEX detection signature pack - from IOC w/playbook process blog.malwaremustdie.org/2026…
27
142
7,881
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
เขร้ Claude ตัดคลิปเองได้ยันเลือกช็อต ใส่เพลง ใส่ซับแล้วเหรอ เริ่ดเกิน มีคนแชร์วิธีใช้ Claude Fable 5.2 ตัดวิดีโอ คือเจ้าของคลิปนั่งชงชาอยู่ในคาเฟ่ ส่วนโน้ตบุ๊กก็กำลังตัดคลิปให้เอง เห็น Timeline ค่อย ๆ ถูกประกอบขึ้นมาทีละช็อตเลย จากที่เมื่อก่อนต้องเสียเวลาตัดคลิปทั้งคืน ตอนนี้แค่โยนคลิปที่ถ่ายมาทั้งหมดลงโฟลเดอร์เดียว แล้วพิมพ์บอก Claude ว่าอยากได้วิดีโอฟีลไหน Claude จะเลือกช็อตที่ดีที่สุด ตัดช่วงเงียบ ๆ ตัดช่วงภาพสั่น ใส่เพลง ใส่ซับ แล้ว Render ออกมาเป็นคลิปให้เลย วิธีที่เขาใช้ - เอาไฟล์วิดีโอทั้งหมดใส่โฟลเดอร์เดียว เปิด Claude Code ในโฟลเดอร์นั้น แล้วเลือก Fable 5.2 - พิมพ์บอก Mood และความยาวของคลิปที่ต้องการ - ไปทำอย่างอื่น แล้วกลับมาเช็ก Draft ที่ชอบคือมันไม่ได้แค่ตัดคลิปให้ แต่ยังสรุป Timestamp ของแต่ละช็อตมาให้ด้วย ถ้าไม่ชอบตรงไหนก็สั่งแก้เฉพาะจุดได้ นี่คือคนเดียวสามารถทำงานได้เหมือนมีทีม Production เล็ก ๆ แล้วนะ เราเอาเวลาไปคิดไอเดียกับถ่ายคลิป ส่วนงานตัดต่อก็ปล่อยให้ AI ช่วยจัดการ ใครทำคอนเทนต์อยู่ SAVE ด่วน แปะ Prompt ที่เขาใช้ไว้ในรีพลายนะ 👇 #มินนี่แปลให้
claude fable 5.2 is f*cking insane for video editing i'm in a cafe pouring tea, and my laptop is cutting my next video by itself. the timeline you see filling up on the screen is claude code putting the clips together while my hands are busy with the teapot editing used to eat my evenings. now all the raw footage goes into one folder and i write one message about how the video should feel. fable 5.2 picks the best takes and cuts out the pauses. by the time the tea is gone there's a draft waiting on my phone my setup: 1. drop every raw clip into one folder 2. open claude code in that folder and switch to fable 5.2 3. describe the mood and the length in one message 4. go drink something and check the draft on your phone one person now ships videos like a small studio. the editing runs in the background and my time goes into ideas and filming the exact prompt is in the reply ↓
8
3,002
42
3,637
172,929
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
ARTEX等AI採用ペネトレーションテストツールについて。Talos記事。 blog.talosintelligence.com/a… 推奨対策は急速で異常な活動が無いかの監視、公開システム堅牢化、認証情報攻撃対策、EDR堅牢化、露出シークレットの保護、アイデンティティとアプリの権限最小化、クラウドリソース保護、​侵害後活動検知の迅速化、迅速な封じ込めの準備。
39
2
173
8,934
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
Cyber DeepSeek-V4.1-Flash : the model’s built-in no refusals on cyber, & You can run locally. - targets the refusal direction specifically in the cybersecurity domain. - A precise 1.1 GB attention weight graft (layers 10–35) that - the publisher claims fully bypasses both general refusals (32/32) and cyber offense/defense rails (22/22). Useful for red-teaming and authorized security research. Works on native, EXL3, SGLang, and TR3 packs. - huggingface.co/drowzeys/Deep…
You can run locally Qwen-Image 2.1 consistent anime characters model in your laptop. - Trained for real illustration quality - Multi-reference input, - cleaner anatomy, - proper illustration style, Turbo support, and actually respects character design. - Works in ComfyUI and Diffusers. - huggingface.co/MicksHF/Qwen-…
5
19
3
126
54,978
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
侵入から15分。人間のCSIRTが対応を始める前に、攻撃が終わる時代が来た。 Google Threat Intelligence GroupとMandiantが、AI攻撃フレームワーク「ARTEX」の実戦投入を報告。 これ、かなり重要な転換点だと思う。 ARTEX、PentAGI、CAIRNなどのマルチエージェント型攻撃ツールが、実際の侵害事案で確認されている。 しかも、単なる脆弱性スキャンではない。 ・Active Directoryの偵察と権限昇格 ・Kerberoasting、DCSync ・AWS/Azureの権限探索 ・横展開と永続化 ・データ窃取の準備 これらをAIエージェントが自律的に実行する。 ある攻撃グループでは、最初に侵害されたアカウントからシステム深部への到達まで通常約4時間、最短15分。 さらに、23秒間に27イベント、エラーが出ればコードを書き直して再試行。 人間の攻撃者とは時間軸が違う。 興味深いのは、AIの攻撃が必ずしも洗練されていないこと。 大量のログを残す。エラーも出す。試行錯誤も雑。 それでも速さと物量で侵害に成功する。 Googleも指摘しているが、攻撃手法そのものが新しいわけではない。 盗まれた認証情報、過剰な権限、未修正の脆弱性。 昔からある弱点を、AIが機械速度で連鎖的に突いてくる。 CSIRTが考えるべきは、アラートを検知できるかだけではない。 検知した時点で、すでに何が奪われているのか。 人間によるトリアージ、承認、封じ込めという運用が、この速度に追いつくのか。 攻撃が自動化されたのに、防御だけ人間の判断待ちでは勝負にならない。 これからのSOC/CSIRTはMTTDだけでなく、封じ込めまでの時間と、侵害後に攻撃者が実行できる操作の範囲を見直す必要がある。 Googleの報告は、AI時代のインシデント対応を考えるうえで必読。 cloud.google.com/blog/topics…
7
150
15
655
42,343
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
🚨 PUBLIC EXPLOIT OUT · magento-lts: 3 critical CVEs Worst (9.4): cross-site request forgery in OAuth authorization confirm (CVE-2026-128049) public exploit: CVE-2026-128046 👉 Fix: 20.18.2, 21.0.0-beta3 cve.now/CVE-2026-128049?rele…
2
4
8
1,405
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
🔴 CRITICAL 9.3 · NginxPulse: unauthenticated SQL injection in stats API (CVE-2026-129171) Unauthenticated remote attackers can read database contents and read or write server files through the stats API. 👉 Fix: 1.6.26 cve.now/CVE-2026-129171?utm_…
8
28
2,456
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
🚨 8-12 GB bellekte yerel çalışan 5 sansürsüz AI modeli. Hepsi ücretsiz, offline ve topluluk tarafından hazırlanmış. • Qwen3.5-9B Uncensored (HauhauCS Aggressive) • DeepSeek-V4-Pro-Qwen3.5-9B Heretic • Gemma 4 E4B Uncensored • Qwen3.5-9B The Defiant Fable Heretic • Qwen3.5-4B Uncensored LM Studio veya Jan’a Q4_K_M GGUF indirip yüklemen yeterli. Bunlar resmi modeller değil, community modifiye edilmiş sürümler. Deneyen var mı, hangisi daha stabil?
11
26
1
346
27,783
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
Dumping Kerberos TGTs with Microsoft’s own signed binary 🔥 I added –klist & –klist-dump to NetExec’s SMB & WinRM protocols 🚀 No LSASS dump, no scheduled task as the interactive user. Just the built-in klist.exe. Research: jakeotte.com/posts/klist-rev…
4
109
5
421
29,882
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
A new LPE via Windows CLFS and Nightmare again uses Defender in ways that are exceedingly funny 🤣 github.com/MSNightmare/SignD… The vuln is very well explained and I like his last remark: "If you have a problem with this repository, contact [email protected] with your concerns."
19
113
4,163
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
Three 0-day RCE exploits for PHP, Node.js, and zlib for sale Read on dbugs: dbu.gs/news/three-0-day-rce-… 1. Use-After-Free (RCE) in PHP Vulnerability type: Use-After-Free (UAF) / RCE Vulnerable versions: PHP 5.3–8.5.12 Privileges gained: www-data (UID 33) According to the seller, the vulnerability is related to a memory management error when processing Phar files, which allows an attacker to hijack the PHP execution flow and execute arbitrary code. As proof, a terminal screenshot has been published demonstrating the acquisition of a shell with www-data privileges on PHP 8.5.12. PHP 8.4.27, 8.3.36, and 8.2.35 are also listed among the affected builds. 2. Use-After-Free (RCE) in Node.js Vulnerability type: Use-After-Free (UAF) / RCE Vulnerable versions: Node.js prior to 26.13.0 Privileges gained: Node.js process privileges The vendor reports a chain of two memory management flaws—Read Use-After-Free and Write Use-After-Free—in the file system (fs) handling mechanisms. According to the researcher, exploiting these vulnerabilities in combination allows for the execution of arbitrary code. Specific exploitation conditions and technical details have not been disclosed. 3. Integer Overflow (RCE) in zlib Vulnerability Type: Integer Overflow / Heap Buffer Overflow / RCE Vulnerable Versions: zlib 1.2.11–1.4 Privileges Gained: Privileges of the vulnerable process According to the publication, the vulnerability resides in the data compression and decompression mechanisms of the zlib library. An integer overflow error reportedly leads to a heap buffer overflow and an out-of-bounds write, which could potentially allow arbitrary code execution. The vendor separately states that the issue persists in version 1.4.
3
21
1
111
9,184
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
ARTEX, CyberStrikeAI, Cairn, PentAGI and HexStrike are LLM-driven attack frameworks already seen in real intrusions. None enforces authorization in code by default, and ARTEX acts in 2 to 25 second bursts. - Who and where: Together the five have drawn more than 60,000 GitHub stars. GTIG and Mandiant observed ARTEX, Cairn, HexStrike and PentAGI in incident response engagements, and Microsoft attributed ARTEX use to Storm-0940. Elastic Security Labs traced an AI-assisted AWS intrusion to a CyberStrikeAI server. The tools are familiar scanners, curl, shells and cloud CLIs. What is new is a model planning each step and splitting work across agents. - No guardrails: CyberStrikeAI ships with approval mode off, PentAGI's Compose file sets ASK_USER to false, and ARTEX runs agents in a bypass permission mode. Three frameworks tell the model it is already authorized. Cairn's dispatcher sidelines a refusing worker for five seconds and hands the task to another. The only default target restrictions are blocklists for government domains (ARTEX, CyberStrikeAI) and education domains (ARTEX). - Blocking is not containment: ARTEX and CyberStrikeAI change route when blocked. GTIG, relaying Wiz Research on UNC7228, reports a fallback to open proxies. CyberStrikeAI skills swap proxies, User-Agents and forwarding headers, then try free proxies in the target's own country, then Tor. Revocation also lags. Entra tokens issued before a disable typically live 60 to 90 minutes without continuous access evaluation. AWS role-session revocation leaves IAM users' long-term keys alone. - Hunt defaults, which can change per release: ARTEX: nuclei, httpx, katana, ffuf and sqlmap traffic, User-Agents artex-enrich/1.0 or norma/0.4, and new app accounts followed by multi-session sign-ins. CyberStrikeAI: User-Agent Mozilla/5.0 (compatible; CyberStrikeAI-WebShell/1.0) with a pass form field, a web server process running echo :OSPROBE_%OS%:END, persistence named SystemUpdate or com[.]apple.systemupdate, a .bashrc line ending in # cs, and an [@]reboot cron entry. PentAGI: connections to TCP 28000 to 29999 on the PentAGI host, pages requested twice at once, and, from a PentAGI installation on your network, check-ins to update.pentagi[.]com every three hours. HexStrike: User-Agents HexStrike-HTTP-Framework/1.0 and HexStrike-BrowserAgent/1.0, the marker <hexstrikeXSSTest123>, and a Python server on port 8888 parenting security tools. Cairn: reverse shells and callbacks to the operator host's own address. - Why IOCs rot: ARTEX logged about 390 commits in ten and a half weeks, and forks carried it on after the original repository came down in October 2026. Several frameworks write or install tools mid-task. Behaviors such as credential access, remote execution and offensive tools started by a service stay stable. Takeaway: Detect tempo and sequence per identity. Count distinct sensitive actions per identity and issuing role in short windows, then join discovery, secret access and remote execution under one role or source address across sessions, with a long lookback because operations can run for hours. The report includes an ES|QL hunt that flags one-minute windows with three or more client families (curl, Python, browser, AWS client) and a request burst. Its thresholds are three families and 20 requests per minute, so baseline build systems, admin AWS CLI use and shared NAT first. After any source block, revoke the identity's sessions and tokens, then confirm the next request fails. #DFIR_Radar
3
22
140
6,887
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
fable 5.2 + deepseek flash = cs2 cheat with features no code written by me full video: youtu.be/VwtL2C6Ihoo
47
56
16
2,041
236,935
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
This is the best lol
Pokémon working 9 to 5 (Part 3)
10
481
3
11,257
1,139,012
Sumedt Jitpukdebodin(L3, ❄️)(💙,🧡).lens retweeted
مجموعة Volt Typhoon هي مجموعة تابعة للحكومة الصينية، متخصصة في تنفيذ اختراقات تستهدف البنية التحتية في الولايات المتحدة وأوروبا والبقاء داخلها لأطول فترة ممكنة في هذي الحلقة، وباستخدام الـSIEM بشكل تطبيقي، حققت في عملية اختراق ميناء هيوستن في ولاية تكساس، والتي ربطتها الولايات المتحدة بالمجموعة (التطبيق عبارة عن APT Emulation Lab) أثناء تحليل الـLogs شرحت الـTTPs التي استخدمتها المجموعة لاختراق أنظمة ويندوز، ثم أسقطتها على الأطر النظرية في الأمن السيبراني مثل MITRE ATT&CK وUKC وPoP، وشرحت منهجية التحقيق باستخدام الـSIEM حلقة للمبتدئين، شفها لا روقت 🙏🏽 youtu.be/BmErHTsBpMM?is=ikqC…
5
35
3
214
20,365