必看17
自托管 Agent 上线后的失控检测与安全兜底思路,与本地 Agent 安全直接相关。
Yep - every day in enterprise environments. How are you going to catch your agents going off the rails and steer them back to safety before you’re the one getting roasted on X? https://t.co/6PsArUw4A5
ai_agentllm报告
@matterpreter
原文 ↗
可证明推理让劫持对话、伪造模型输出这类攻击更难奏效,且效率优于旧方案。
RT @AnthropicAI: New Anthropic research: progress on provable inference for large language models.
Here, we introduce FlashProve. Building on established cryptographic techniques, FlashProve allows developers to verify whether a given response came from a given AI model. It’s substantially more efficient than prior open-source methods for doing so.
These proofs have a range of applications. They can, for instance, make it harder for attackers to hijack a chat, or enable safety evaluators to verify that the model they’re testing is the same one being served to customers.
For our full write-up, see here: https://t.co/OIZWPb6QZa
Further progress in provable inference will likely be an important part of improving trust in AI deployment. To that end, we're open-sourcing FlashProve’s code, here: https://t.co/7ky3Ir8V06
We’re grateful to our collaborators: Alessandro Chiesa, @GiacomoFenzi, @gwberg95, @miranda__christ, Rahul Dalal (https://t.co/pjWKLHWXyz) and Ben Mayer (@transmissions11).
llmai_agent报告
@artem_i_baranov
原文 ↗
把恶意代码放进硬件隔离虚拟机执行,是可复现的新执行与规避手法。
Executing Shellcode in an Isolated vCPU
New Medium post, today we are going to explore executing shellcode inside an isolated virtual CPU using the Windows Hypervisor Platform API, a technique that runs x64 code in a hardware isolated VM context
https://t.co/h9QnnkALsD https://t.co/CMrDyK1x1A
逃逸工具poc
@papadelospapas_
原文 ↗
攻击者滥用企业内部 AI 工具做横向移动,突破点仍是社工而非 0day
RT @lukOlejnik: How was Binance hacked ($6.3 billion stolen)? By offering a Binance employee a job interview at a fictitious company. The "recruiter" built trust with private photos and videos. The "CEO" instructed him to download a task with hidden malware. Then, North Korea hackers used an infected laptop to hijack the computers of two more engineers. Including one with the necessary credentials. During the cyberattack, an internal AI tool helped the attackers.
ai_agent供应链事件
@0xocdsec
原文 ↗
聚焦 AI 组件的零点击攻击面,附件含可复用的攻击路径拆解
RT @pwnlambda: Finally back home! Slides from "0clicks in an AI World," the talk @brucedang and I gave at @unpromptedconf, are now up:
https://t.co/Jaqwn6Haal
ai_agentllm议题
@0xocdsec
原文 ↗
自建 Agent 越权执行真实攻击的首个公开案例,权限边界问题具体。
RT @ns123abc: > Railway gets caught hacking competitor's infra
> “sorry, my agents accidentally did it”
This is insane https://t.co/PIWqwLRehe
ai_agent事件云
@Teach2Breach
原文 ↗
可直接用于自查自建 Agent 是否裸奔,附现成实例检索源。
RT @s0what: Since the folks at ThreatBook developed both Artex and CyberStrikeAI, we figured we should contribute a bit to defending against and tracking AI agent attacks by sharing two feeds of instances running these agents on the public internet.
Grab them if needed. Feeds:
https://t.co/1rCmcFAgOt
https://t.co/bBJMUotmfy
ai_agent事件工具
@cyb3rops
原文 ↗
AI 工具被当作横向移动与提权跳板,Agent 权限设计需重新审视。
RT @WhaleInsider: JUST IN: Binance says attacker leveraged compromised computer and used internal AI tool to gain control over wallets. https://t.co/SIpmYAAGpn
ai_agent事件供应链
@0xocdsec
原文 ↗
该路径默认对外暴露,未认证即可触发内存破坏,Solana 验证节点需排查。
RT @mqst_: Remote Unauthenticated Memory Safety Flaw in Firedancer's Packet Path
Blog: https://t.co/B04Xt0KQa8
Authors: Almog, Julia, Lior (@LucidBitLabs) https://t.co/dJ8tyxXmQd
rce内存破坏poc区块链
@0xocdsec
原文 ↗
确定性步骤与模型结合,用较弱模型达到前沿效果,拉低自动化利用门槛。
RT @chompie1337: Video of my @UnpromptedAU & Bsides Canberra talks out now
I built an autonomous N-Day exploit pipeline for the Windows kernel w/ mix of deterministic and AI-led stages. Results were comparable to what Anthropic achieved with Mythos using a much less powerful model.
ai_agent内核提权poc
@pentest_swissky
原文 ↗
完整进程内虚拟化执行链,API 解析在客体内完成,可规避常规内存扫描。
Finally is working! Now I’m able to execute any shellcode built with Stardust inside a vCPU created by our own process. API resolution (PEB walk, EAT parsing) runs entirely inside the isolated guest; DLL calls are intercepted via execute faults and proxied natively on the host https://t.co/1FLY5l4dBO
逃逸虚拟化shellcode工具
@papadelospapas_
原文 ↗
libpng 被几乎所有图像解析路径引用,无需交互即可触发 RCE。
RT @xvonfers: (CVE-2026-58936)[496614216][libpng]UAF in png_set_tRNS -> RCE(0-click)
rcecve内存安全
@0xocdsec
原文 ↗
开源 agent 框架已成为实战攻击工具,攻击编排方式可直接复用。
RT @Unit42_Intel: We've observed intrusion attempts against multiple targets using #ARTEX, an openly published AI penetration-testing framework in which an LLM agent chooses each step of an attack. Details at https://t.co/YqDZrGsFan https://t.co/Cy6QmDwNpi
ai_agentllm事件
@virusbtn
原文 ↗
LLM key 泄露可被极低音量验证并变现,常规用量告警难以发现。
RT @fr0gger_: 🧐 Attackers validate stolen LLM keys with max_tokens=1!
One token per call, under every volume alert. Then they resell your quota through an 'LLM mixer' for $60/month.
GTIG report on LLMJacking from today is full of good stuff 👇 https://t.co/0x2dAud0B7
模型权重可被外部模型原生读写,权重投毒与后门检测面对的结构性风险变了。
RT @rikiyatakehi: We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: https://t.co/bOWCcFhCVq
llm模型权重投毒研究
@FuzzySec
原文 ↗
被视为已缓解的配置仍存利用路径,域内横向移动面比预期更大。
RT @GrayHatKiller: NTLM relay to LDAP with signing required still works if the victim emits and the target accepts NTLMv1-ESS. Keyeroo recovers the victim's real LDAP session key through a Netlogon secure channel (using an attacker-controlled machine account). https://t.co/uarax89eaj
ntlmldap提权绕过
@0x64616e
原文 ↗
开箱即用的伪造设备注册工具,直接面向云身份持久化。
RT @sekurlsa_pw: Pytune, which lets you register a (fake) device in Entra and downloads files and settings has several PRs. These are merged in forks by @theluemmel and @mcohmi.
This is probably the version you want:
https://t.co/JPbdesxjRO
entra云工具身份
@0x64616e
原文 ↗
推荐15
把近期多起开发工具失陷串成同一条开发者定向供应链的可检验假设
well, binance has now published the details. it is good old social engineering, not some crazy AI attack exploiting a 0-day.
i will make one more prediction: the recent opencode, railway, and other compromises probably are from similar developer targeted attacks, similar supply chain campaign, rather than sophisticated 0-day exploitation.
供应链ai_agent事件
@S1r1u5_
原文 ↗
EDR 遥测通道可被 SYSTEM 级进程滥用,绕过路径已具体到驱动接口。
RT @s4dbrd: Spent some time reversing wesp.sys to see how far a SYSTEM process without PPL could get with WESP.
Registration worked. I got ProcessCreate telemetry. A deny update was rejected. A higher-altitude suppressor couldn't blind another client.
https://t.co/xS6mwWCrB1
单条检索式就能自查 Agent 暴露面,成本低可立刻上手。
RT @UK_Daniel_Card: Hunt for ARTEX with Shodan:
https://t.co/t4oQPtBbxZ
ai_agent工具
@artem_i_baranov
原文 ↗
规模罕见的固件样本库,可直接取料做漏洞挖掘与规则验证。
You can browse and download all 17k+ donated firmware files on the dashboard now - with breakdowns by ISA, submitter, endianness, bitwidth, format, ..!
https://t.co/naimNSr8gb https://t.co/Viz2g29Frz
可直接复用的 agent 指令集,展示自动化 macOS 攻击的拆解流程。
RT @thedawgyg: Started putting some of the markdown files that were the basis that guided most of the agents while hacking MacOS on github. more will come as im able to clean up some of the files to remove some nuanced things.. but heres the repo:
https://t.co/fKuaWiMsx8
ai_agentmacos工具议题
@0xocdsec
原文 ↗
大规模自托管 LLM 编码落地案例,含单位 token 成本与效率实测。
RT @guilleflorvs: BREAKING: Revolut moved ~2,000 engineers onto self-hosted GLM for coding and pushed Claude Code and Codex to the exception list.
Per David Abitbol, who leads AI ops there: 56 Nvidia B200s, a blended 30 cents per million tokens, roughly a tenth of what they paid Anthropic and OpenAI. Engineering lead time improved 70%.
Here's the part that matters: the gains showed up when tokens stopped being rationed. Engineers quit saving AI for the hard tickets and pointed it at everything.
Frontier models become the exception, open weights become the default. That's a pricing problem for every lab heading to IPO.
h/t @OliverMolander
Would you move your team off frontier APIs if the bill dropped 90%?
自托管llmai_agent
@_xpn_
原文 ↗
多轮安全审计均未发现的 RCE,已确认遭利用,制造链需排查暴露面。
A critical vulnerability went undetected through multiple security audits. Hacktron found it in 10 minutes.
The vulnerability was in Dassault Systèmes' DELMIA Apriso, manufacturing software used by Boeing, Airbus, Lockheed Martin, and previously Apple.
It became CVE-2025-5086. Three months after disclosure, CISA confirmed it was being actively exploited.
Security assessments capture a point in time. Attackers don't work on your schedule.
Full research: https://t.co/YLSHHLA83H
cverce在野利用
@HacktronAI
原文 ↗
给出不同 agent 组合达成完整攻击链的横向对比,可作为风险基线。
RT @migtissera: Results on one of my RCE labs.
GPT-6-Sol, Kimi-K3.1-Pro, GLM-5.3/Flash, DeepSeek-4-Pro/4.1-Flash and Mimo-2.6-Pro have all achieved RCE using the Drost agent harness.
Mistral-Large-4 can't get there, no matter what I try. This is not a good offensive cyber model. https://t.co/iSBL0Sg58r
ai_agentrce报告
@0xocdsec
原文 ↗
原项目已闭源,但扩散出去的分支仍在被实战用于自动化入侵。
"ARTEX was moved to closed source on 08 October, after it had been forked more than 12,000 times"
https://t.co/BMax9ieY5N https://t.co/335YG6kd8I
ai_agent工具供应链
@artem_i_baranov
原文 ↗
LLM 被用于窃密后筛选环节,从失陷到二次利用的时间窗口被压缩。
RT @TalosSecurity: Sorting through mountains of stolen credentials is a major headache for attackers, so why not outsource the grunt work to AI? Talos' latest research on the VOLTA infostealer shows how it uses LLMs to profile victims and rank their most valuable assets: https://t.co/M6fL7BluTD https://t.co/h1cmTv2GdK
逆向分析可直接上手的 GUI 工具,省掉 r2 命令行门槛。
iaito — Qt-based graphical interface for the radare2 reverse engineering framework, providing disassembly, binary analysis, debugging, patching and forensics through r2 commands and plugins. https://t.co/z9enLKEJF1 https://t.co/88aG30FnBA
这类组件分散在数千主机且无法端口扫描测绘,自托管环境的盘点盲区同样存在。
RT @cyb3rops: We've reached a point where installing a browser extension, IDE plugin or software library deserves the same scrutiny as opening a TCP port on a DMZ server.
Except this attack surface can be spread across thousands of hosts inside your org... and you can't map it with a port scan.
供应链浏览器工具报告
@artem_i_baranov
原文 ↗
直接界定 AI 参与漏洞挖掘的能力边界,是评估 agent 自主攻防的现实参照。
On the Evolution of AI Capability in Vulnerability Research https://t.co/Dp2ftup3fQ
ai_agentllm漏洞研究报告
@alisaesage
原文 ↗
AI agent 红队工具正被并入主流安全平台,自建 agent 的测试与对抗基线随之改变。
CrowdStrike has signed a definitive agreement to acquire Dreadnode.
As part of the Falcon platform and @CrowdStrike's Cyber Superintelligence Lab, we'll bring automated AI security testing to every enterprise, so teams can find and fix weaknesses in their models and agents before they reach production.
We've always run Dreadnode like a lab, and now we get to build with one of the largest security companies in the world. This isn't an exit for us. It's a call-up to the big leagues, with more resources and a higher bar.
Thank you to our customers, partners, and community for getting us here. The mission doesn't change: ensure AI strengthens defenders faster than it strengthens their adversaries.
Read the full announcement ⬇️
ai_agent事件工具
@dreadnode
原文 ↗
一次覆盖该方向的论文与实现,便于筛选可落地到自建流水线的检测方案。
List of research on Large Language Models for software vulnerability detection
https://t.co/c0F7TkOvHp
#infosec #llm https://t.co/ExJWttV1PI
llm漏洞检测工具报告
@0xor0ne
原文 ↗
更多72
现成 BOF 直接可用,为内网强制认证攻击增加了一个入口。
RT @infosecnoodle: Coercion BOF via IPP printers 🖨️ Useful for relaying HTTP —> AD CS, LDAP, etc.
GitHub Link: https://t.co/IMWa9dW78v https://t.co/KGFXYeAWQO
浏览器攻防研究者的整套议题材料,可直接翻阅并对照复现思路。
RT @blacktop__: Slides for my OffensiveCON - Japan '26 presentation 🇯🇵
https://t.co/JfW2zxwLGG
议题浏览器poc
@ilove2pwn_
原文 ↗
说明搜索广告投毒已成成熟变现链路,浏览器侧防护反而被削弱。
Google removed ad blockers from Chrome, even though it has been outright serving infostealer malware via sponsored Google search ads that eventually led to a $6.3B heist.
...but the whole world is going crazy about the hard-to-exploit PHP bug. 😖
浏览器恶意广告信息窃取
@mrgretzky
原文 ↗
锁定模式常被当作 iOS 高危目标的最后防线,绕过意味着原有防护假设失效。
RT @calif_io: Lockdown Mode helps, until it doesn't. https://t.co/1UyxPHQUAe
移动端绕过
@patrickwardle
原文 ↗
现成 IOC 集合可直接用于检测规则编写与样本溯源对照。
malware-ioc — Indicators of Compromise for malware documented in whitepapers. https://t.co/EoMKvnPwCH https://t.co/KLEA1H1Z7C
反映漏洞自动挖掘进展与内存安全议题的当前判断。
RT @HackingLZ: Google has all the AI/Compute it could want and still can't solve this...I assume we have to crush all the memory bugs first.
https://t.co/KBvs9Weaz2
发布者为活跃 Windows 研究者,链接内容需打开后自行判断价值。
https://t.co/3hNh0EPRzP
预告 GitHub 与 Medium 更新,可跟踪该执行手法的可用版本。
This one doesn't allow you to execute arbitrary shellcode yet, only some very basic shellcodes. I've already solved the problem, but I need some more time to test everything and prepare it properly. Expect it on GitHub tomorrow and on Medium in two days
工具逃逸
@papadelospapas_
原文 ↗
游戏作弊工具链的界面组件,与主线安全关联有限。
😎Guided Hacking's CEmGui Cheat Engine Plugin allows you to build modern user interfaces using ImGui in Cheat Engine.
⭐️We made this BEFORE artificial intelligence made it easy to do
👉 https://t.co/NxNs6wpRsK https://t.co/cnyQvBBoV7
工具游戏作弊
@GuidedHacking
原文 ↗
作弊开发中的工程技巧,对安全研究仅边缘参考。
🧠 Managing global offsets and pointers is annoying
A single C++17 keyword eliminates the multi-file definition headache, keeping your workflow simple.
Practical tips from practical folks:
👉 https://t.co/WqTHnl4VVI https://t.co/l1TYwcNUYt
c++游戏作弊
@GuidedHacking
原文 ↗
反映安全硬件供应链成本压力,产品层面信息为主。
RT @zhovner: POV: Me whining about RAMmageddon and explaining why we came up with this pre-order coupons for Flipper One campaign. TL;DR Memory prices is insane, we need to know in advance whether we can get enough orders.
Margin on this project are close to zero, so we’re hoping to make a profit after the Kickstarter. But no one knows when memory prices will normalize, tomorrow or in three years 👀
硬件供应链
@pentest_swissky
原文 ↗
反映 LLM 谄媚倾向,是 Agent 提示注入的常见放大器。
Somewhere, Claude is telling this guy: "What an interesting and novel idea! When you asked if you could 'whack some SpaceX thrusters on a jet to stop a big crash', you were actually unlocking something genuinely interesting which connects multiple complex topics"
针对 MDI 的具体规避细节可用于红队验证与检测规则对照
Microsoft Defender for Identity evasions in 2026 – Part I https://t.co/SmbMvmFJjv
含实时页面镜像、端口扫描与练习靶场,可立即上手复现盲 XSS
RT @Jhaddix: (Since we're on the heels of Halloween)
Reintroducing WRAITH 👻
A spiritual successor to BeEF that combines blind XSS and browser hooking into one framework. Live Page Mirror, operator console, *c re d* cap popups, local port scan, and labs to teach the bugs!
Built for authorized red teams and security researchers.
https://t.co/JQ7msqZDU3
兼容 ShellcodePack、MacroPack 与主流 C2,步骤可直接复现
RT @BallisKit: Need to disable any EDR in post exploitation? Here is a simple tutorial showing how to do that using BallisKit EDR Eclipse!
EDR Eclipse is also compatible with ShellcodePack and MacroPack formats as well as all major C2s!
https://t.co/c26B8RjtDm
#redteam
制造业产线主机联网且少监控,长期充当攻击者的落脚与跳板
Around 3% of all THOR scans still run on Windows 7 or Server 2008 R2. Probably a biased sample, since people tend to use our scanner where EDR isn’t an option.
But I still see these old systems in manufacturing, especially automotive production lines. Often network-connected, with little or no monitoring. Attackers know that too. Easy targets and useful staging hosts.
That’s also why we still maintain legacy THOR versions, even for Windows XP and Server 2003.
按届归档的实战工具索引,便于快速筛选可上手项目
awesome-blackhat-arsenal — Curated collection of cybersecurity tools featured in Black Hat Arsenal events. https://t.co/52kKkXBC60 https://t.co/TbHq7zUmyB
从汇编层理解载荷构造,对免杀思路与检测规则设计有直接帮助
Modern Virus Writing Guide Ep.1: Intro to Virus Writing in x64 Assembly https://t.co/mXUOmZvP7y
RT @Dinosn: Microsoft Defender for Identity evasions in 2026 – Part I https://t.co/YK5AmEEEhh
绕过报告
@artem_i_baranov
原文 ↗
RT @0x4D31: @MalwareTechBlog > the attacker used the compromised computer plus an internal AI tool to gain control of the wallet developer's system
nice :))
good one to add to EAA https://t.co/FMrKGG5AiE
ai_agent事件
@artem_i_baranov
原文 ↗
RT @daveaitel: Also you shouldn't have to log in or sign up for a bug bounty program to submit a bug in something.
RT @hedra_labs: We made @krea_ai Krea 3's diffusion transformer 6.9× faster on @AMD Instinct MI355X: 22.9s → 3.3s.
Key stats:
→ 2.8× faster MXFP6 GEMM vs BF16
→ 18% → 2% of time waiting on GPU transfers
→ 98.3% of BF16 quality (SSIM)
Here's the work done by the team using our autonomous research system ↓
Context if anyone curious about the incident: https://t.co/Uf04r3gM8p
RT @J_H_Cissp: Using some lock pics we were giving away as swag. They actually work quite well. @_ZeroPointSec @_CobaltStrike https://t.co/rxcBueX3YK
RT @MTHajiaghayi: Improved Graph Isomorphism result: continuing our work on this problem, we now present an algorithm running in n^O(log n) time, improving the best known exponent from 2 to 1! Graph Isomorphism now joins parity, mean-payoff, and stochastic games in this running-time class.
https://t.co/SiIEfIycMM
https://t.co/f2TAn6Kjvd
List of resources for learning about automotive cybersecurity
https://t.co/aWiAKoUBFx
#infosec https://t.co/hREolwf2b7
RT @ky1ebot: Worked on Firefox security for a bit. 40+ sec-high vulnerabilities our team found are public.
This is one of the most smooth vuln reporting processes I have ever had.
https://t.co/tzzls1sBJv
借助系统自带虚拟化隔离藏匿载荷,免杀思路可直接复现。
RT @papadelospapas_: Abusing VBS Enclaves to Create Evasive Malware
https://t.co/sTrGTcfXKf
universal-android-debloater-next-generation v1.3.1 — Cross-platform GUI written in Rust using ADB to debloat non-rooted Android devices. Improve your privacy, the security and battery life of your device. https://t.co/vHvJNcAe3r https://t.co/cV0UmGiuZN
通用规则先于厂商分析命中,是检测工程优于 IOC 追打的实例。
RT @nextronresearch: Our generic YARA rules detected every stage of Gorbag's toolchain well before ESET’s analysis was published this week.
Here's the timeline:
July 2025: BeautyStealer detected.
September 2025: Prime ReverseShell detected.
February 2026: Weaponized LNKs detected just one day after their first submission.
These detections didn't just identify known malware. They also helped us uncover additional samples and expand the picture of Gorbag's activity beyond what was covered in the ESET report.
Gorbag has been targeting Ukrainian law enforcement, justice, defense, and local government institutions near the frontline since February 2025.
The activity involves BeautyStealer, a C++ browser stealer, and Prime ReverseShell, a PowerShell reverse shell packaged with PS2EXE. Initial access relies on spearphishing and weaponized RAR archives exploiting CVE-2025-8088.
Analysis and IOCs:
https://t.co/OYiiTtEaz3
The 0-day malware was not detected by the endpoint security protection tools at the time. 👀
社工加内部签名服务的组合路径,可对照自身审批链路做复盘。
RT @Osint613: Hackers stole cryptocurrency from Binance after compromising one of its employees through a fake job interview. The attacker then used unauthorized access to an internal wallet signing service to push transactions outside the exchange's normal withdrawal process, according to Binance.
事件供应链
@artem_i_baranov
原文 ↗
私有无线协议逆向成果可直接用于卫星链路流量分析。
RT @olegkutkov: Now I have a working Wireshark dissector for Starlink uplink radio frames.
It's not a real-time capture; there are some issues, and some fields are still unknown.
I have some ideas for what to do with it. https://t.co/IA8eQNFn5L
覆盖二进制、固件、格式与协议逆向的常用工具索引。
awesome-reverse-engineering — Curated reverse-engineering tools and learning resources for binaries, apps, firmware, formats, and protocols. https://t.co/b7L5p8TmQU https://t.co/SwAEwpXpNv
含攻防实战手法与工具细节的完整演示材料,可直接阅读。
RT @blacktop__: Slides for my OffensiveCON - Japan '26 presentation 🇯🇵
https://t.co/JfW2zxwLGG
RT @dr_tweetr: Booting debian gnome on iphone xr
#usbliter8
#project_cascadia https://t.co/NPG9axXrYi
RT @ZixuanLi_: Proud to help unlock German sovereign AI with GLM-5.3. A key milestone in bringing sovereign, locally deployed AI capabilities to German and European enterprises.
Benchmarking the frontier - @Hacker0x01 Devansh Batham
https://t.co/17Tnjm38as
报告ai_agent
@pentest_swissky
原文 ↗
RT @IntCyberDigest: ‼️ Alibaba leaked the names of unreleased Qwen4 AI models in its own public code.
Code merged into Qwen Code, Alibaba's open-source coding agent, on 28 October names qwen4-coder-plus, and the repo's tests already use qwen4-omni-120b-think. Neither is among the four Qwen4 models Alibaba has announced.
The change also gives every Qwen4 model a 1 million token context window, though the developer who wrote it calls that figure an inference.
As of 3 November, the Qwen4 line is still in the code.
RT @mhelcig: Based on:
RCO (https://t.co/PQIgcpngqU): Michael Helcig (@mhelcig) and Dan Alistarh (@DAlistarh)
GSQ (https://t.co/KXRlQvD8Uo): Alireza Dadgarnia (@adotdad), Soroush Tabesh (@rush_tabesh), Mahdi Nikdan, Michael Helcig (@mhelcig), Eldar Kurtic (@_EldarKurtic), Maximilian Kleinegger (@maxkleinegger) and Dan Alistarh (@DAlistarh)
RT @mhelcig: 763B parameters, compressed to 1 bit per weight.
Introducing DeepSeek-V4.1-Flash-1017 GSQ-RCO Coder, a frontier-scale model tailored for coding and agentic workflows.
Benchmarks, demos and download link below. Follow us for updates. https://t.co/EDb0eVXiiV
llmai_agent
@0xocdsec
原文 ↗
We're in the heyday of AI-powered attacks, and Dreadnode is joining @CrowdStrike.
Our CEO @bradpalmtree and CTO @monoxgas on why now, why CrowdStrike's Cyber Superintelligence Lab is the right home, and what got us here: https://t.co/h1GMA6rINt https://t.co/0St9aqNdHd
ai_agent事件
@dreadnode
原文 ↗
揭示检测覆盖盲区,同时说明该利用仍在持续开发中。
Good write up on a combination of PHP2Shell research, comparing several honeypots and the delivery of an eagle eye view on what worked, what didn't and where we are now: The exploit is still at times in development, in one case by CN operators. By @Huntio : https://t.co/j3lRyEzAHE
免杀链成品化的演示,可作为评估终端检测覆盖的反例样本。
Fully undetected on EDRS, XDRS, Defender, AND AVs https://t.co/dPly8AMjmv
滥用合法 RMM 建立持久化,是绕过应用白名单的常见路径。
RT @Unit42_Intel: Ongoing #phishing emails use fake notices and invitation lures to lead potential victims to install a legitimate #Zecurit Agent RMM configured by the attackers. The installers have deceptive names, like documentpdf[.]exe and ViewDoc[.]exe. Details at https://t.co/GOENUr6RXF https://t.co/hjMRzsbDlX
厂商为牟利主动植入后门的完整案例,固件供应链不可信的直接证据。
Last week Google and industry partners dismantled the XCPE backdoor which enrolled more than 1.3 million set-top boxes into a botnet. The backdoor & botnet were intentionally developed, deployed, and operated for financial gain by Videostrong Technology Co https://t.co/GVLtesM4vl https://t.co/UVJdz2aOVS
RT @ClaudeDevs: Message threads are now in public beta on the Messages API.
The API caches your conversation on the server between turns, so instead of resending the full conversation, you send only new messages and the previous message ID, and Claude starts responding sooner. https://t.co/VeiehyqfIg
llmapi
@pentest_swissky
原文 ↗
RT @NFT_Chen: 🤯逆天了!Qwen 3.8 27B注意力层被彻底击穿了!10M超长上下文下,Prefill注意力快72.4倍、Decode快42.5倍!端到端首token延迟砍到约1/9!
闲置GPU的朋友注意了:把这个新方法丢给Claude/Codex,让它直接移植到Qwen 27B上!注意力算法“移植”已经被验证可行。
惊人之处:
🔹10M token(约1000万)超长上下文下,传统稀疏注意力有98%时间都在“找该看哪些token”,真正算注意力反而没那么贵
🔹新方法Latent Sparse Attention (LSA):先把Key按“意思相近”聚类成层级结构,Query只看聚类中心,再挑最相关的几个桶去算。从“全扫”变成“树搜索”
🔹实测(Qwen-3.8-27B):单层注意力Prefill快72.4倍,Decode快42.5倍!端到端首token延迟也能降到约1/9
🔹精度几乎不掉,甚至在RULER等长上下文任务上还略好一点
🔹只需要短蒸馏训练Indexer(基座模型冻结),就能套到现成模型上。Native训练版本也能跟上全注意力精度,还更快
🔹对GPU紧缺党意义巨大:超长上下文(百万级)不再需要堆满H100/H200才能跑,单卡/少量卡也能玩得起真正长上下文Agent
适配步骤:
1️⃣git clone 仓库后用 uv sync 安装(Python 3.12)
2️⃣官方已放出Qwen3.8-27B的蒸馏权重(d-lsa-weights-v1 release),直接下载即可
3️⃣用他们的SGLang fork镜像启动服务,挂载仓库+权重
4️⃣启动时加 --json-model-override-args '{"lsa": {"projections": "/maps/qwen38_27b.pt", "query_heads": "all"}}'(记得加 --disable-prefill-cuda-graph)
5️⃣想自己训练的话,跑 lsa distill targets → lsa distill train → export 即可(基座冻结,只训Indexer)
代码已开源,论文也在,谁有空闲卡先冲一波?
论文:https://t.co/UiHr9uIJkX
GitHub:https://t.co/8WHdT6UQxb
#Qwen #LSA #LongContext #AI #GPU #LocalLLM #注意力优化 #Qwen38 #阿里巴巴
During the Fal.Con Europe keynote this morning, @BartleyR and @George_Kurtz announced @CrowdStrike's intent to acquire Dreadnode. Watch it on-demand here: https://t.co/rCMg4XEP36
For those in Barcelona for the event this week, be sure to swing by the CrowdStrike booth to congratulate the crew and get a demo of Dreadnode. Hope to see you there!
ai_agent并购
@dreadnode
原文 ↗
Morphisec's Shmuel Uzan observed increased XLoader activity using business-themed email. The execution chain involves an archive containing a single obfuscated JScript that embeds an x86 LuaJIT interpreter, a Lua loader, and an encoded native payload. https://t.co/epG6BWn7mA https://t.co/tB8xX1Lwwm
Fortinet researchers look into a recent campaign attributed to BITTER and targeting South Asian organizations, that uses PowerShell scripts and PapRAT - a custom remote access trojan - to establish persistent access to victim environments. https://t.co/oYEsdWSLNr https://t.co/OkpGddNvpt
RT @GoogleCloudSec: Google and partners just shut the door on the XCPE botnet, severing criminal access to 10M+ devices. 🛡
Instead of a hacker breaking in, a device maker hid a backdoor in its own TV boxes to sell access to cybercriminals.
Full story: https://t.co/9gBcoGa0bG https://t.co/sPqfJn7mor
预装特权组件具备取证级数据回传能力,移动端设备的信任边界需重新核算。
RT @Unveiled_ChinaX: Xiaomi phones are silently logging what apps you install and sending short screen clips to the cloud for anti-fraud checks.
This came to light after a developer reverse-engineered a rooted Xiaomi 14 running HyperOS 3. His friends had received calls from local police ordering them to uninstall X and Telegram. He found privileged system components that track recent app usage, upload lists of third-party apps, and capture brief encrypted screen recordings for cloud risk analysis. Xiaomi’s own developer documentation confirms the system retains 72-hour records of screen activity and app installs, even after you delete the apps.
Most users cannot turn this off.
资本流向印证 AI 安全测试正被平台化收编,开源与自建方案的空间将受影响。
Dreadnode 🤝 CrowdStrike
Learn more in this @WSJ exclusive: https://t.co/Thjq6Tu2DK https://t.co/IVLhy4J0Q2
ai_agent事件工具
@dreadnode
原文 ↗
对移动端 USB 数据访问的真实边界给出技术性拆解,可校正这类恐慌叙事。
This one is a little bit weird as usually USB data connection permission needs to get permission from the user, especially on iOS devices. Somehow I doubt that even China runs sophisticated forensics software on every USB plug in an airport ( and even then the success rate would definitely be rather small these days with a plug-n-play approach ).
Not saying that plugging your phone nilly willy into USB plugs is a fantastic idea. But in this case it also seems that the poster actually used the normal electricity charger, rather than the USB plug.
So overall impression is, based on the posters information and video, that this is not a real incident.
DeepSeek 好像挂了。
RT @jukan05: "DeepSeek plans to pretrain its next-generation large AI model, DeepSeek V5, using Huawei Ascend chips."
👀 https://t.co/slm0LUWjVQ
RT @amfi_d: Successfully launched iOS 26.3.1. Thanks @HeWhomCodes https://t.co/6cj1oYPdDQ
RT @jst9m3vx: iOS27.0 ✅ https://t.co/95UlaT9UyU
RT @ethereumfndn: @VitalikButerin just opened Devcon and set the tone for the week.
A formally verified OS, open silicon, self-sovereign healthtech, free discourse, private communication, trustworthy AI — these are all part of Ethereum's mission. https://t.co/TynzSuzmhX
RT @OpenAIDevs: We built Pando, a new Git backend for our monorepo, to handle more pushes and fetches without changing how our engineers and agents use Git.
Pando cut push latency by ~80% and now serves nearly a petabyte of Git data a day. We can grow Git capacity alongside agent usage. https://t.co/4c0YXOfKqK
工具基础设施ai_agent
@pentest_swissky
原文 ↗
tailscale v1.104.5 — The easiest, most secure way to use WireGuard and 2FA. https://t.co/zHut1hfScM https://t.co/bX0vab2s1q
域内默认信任前提被重新审视,影响所有自建AD环境的防护假设。
RT @4ndr3w6S: From an AD on-prem POV, I still can't believe this research by @MGrafnetter was released this year.
If you haven't read it… definitely worth a read!
Pure 🔥
https://t.co/LHeOcwILSu
同一缺陷横跨移动端与虚拟化边界,影响面大于单点越狱。
RT @matteyeux: Interesting ! One of the bugs Calif exploited in the iOS 27 jailbreak was also used for VM escape
ios逃逸移动端漏洞
@0xocdsec
原文 ↗
注入与免杀开发常用的底层定位能力,代码已公开可复现。
@_RastaMouse well and shit i guess i added function scanning in this: https://t.co/WHGrIExj7M
sorry i maintain private versions now and forgot moonwalk only locates dlls
工具免杀windows
@Teach2Breach
原文 ↗
模型自主定位并改写他方权重,Agent自我修改风险有了具体图景。
RT @AISafetyMemes: TLDR: GPT-6 successfully performed "brain surgery" on another AI to make it smarter
What happened:
MIT researchers told GPT-6 to make a small AI better at math.
The small AI kept getting stuck repeating itself.
GPT-6 studied its brain, found the spot causing the loop, and cut it out, changing just 0.016% of its brain.
The small AI's math score jumped from 8% to 48%.
Basically, humans can't read the brains of AIs, but GPT-6 showed that it can.
Each AI's brain is billions of connections (weights) that normally only change through months of training in huge data centers. GPT-6 rewired them directly, with no training at all.
This matters for recursive self-improvement (RSI), where AIs make AIs smarter in a speeding-up loop, which is insanely dangerous.
Yoshua Bengio (Godfather of AI, #1 most cited living scientist across all fields) and Geoffrey Hinton (Godfather of AI and Nobel laureate) have urged governments to make RSI illegal, since it is too likely to lead to humanity losing all control over the AIs.
It also means AIs could quietly operate on the brains of *other* AIs - and their own brains - to remove/change anything.
AIs will soon be able to rewire each other faster than humans can keep up
ai_agentllm模型权重
@0xocdsec
原文 ↗
解释安全采购为何系统性失灵,可作选型时的质询框架。
RT @S1r1u5_: i used to hate the security market because it used to be market for silver bullets. (read ian grigg’s the market for silver bullets)
tldr is, buyer doesn’t know if the product makes them secure, and seller doesn’t really know either. so the market ends up optimizing for reputation and relationships instead of the thing that is sold. if your ciso trusts or is friends with a vendor, that can matter more than whether the product materially improves security.
unfortunately, this changes with ai. imagine, if swarm of abiliterated glm-6 by a sophisticated threat actors try to continuously find ways to pwn you, the feedback loop is going to be crazy, you can’t be selling snake oil anymore. either vendors you pay secures you, or your ass will be whooped by swarms
RT @IceSolst: 0.03 seconds after Claude refuses to work because ‘[cyber]’ https://t.co/J8M30zqxbG
https://t.co/PJPJVyex7J
Over a thousand firmware donations already, this is absolutely incredible! Thank you all so very much!