⚡Megatron LLM Hub
All digests
Twitter

推特安全流

Top8

LLM驱动的C2已在真实攻击中落地,自建Agent需评估同类滥用面。

RT @e_kaspersky: CyberStrikeAI in the wild: the use of AI in real-world attacks on organizations We've detected a cyberattack against the education sector in Africa and Latin America where CyberStrikeAI played a key role in the post-exploitation phase. This is an open-source framework with an LLM orchestrator on the command-and-control server plans actions, selects tools, and generates commands, while an implant executes them on compromised hosts. Using CyberStrikeAI, the attackers progressed from reconnaissance to full domain compromise. Read on 👉 https://t.co/29uY3a4diN

ai_agentllm事件 @artem_i_baranov source ↗

针对AI渗透代理的防御新思路,含ARTEX模式与手工部署导出。

RT @Jhaddix: 🛞 Introducing SpikeStrip, a WIP project exploring defenses against AI pentesting automation. Sprinkle reversible prompt decoys, fake vulnerability leads, and math traps into web source. Includes an ARTEX-specific mode and raw export for manual placement. https://t.co/IjURyH0tHs w/ 🫶 from @arcanuminfosec and Codex

ai_agent工具 @0xocdsec source ↗

开源 LLM 编排的 C2 框架首次见于真实后渗透,自建 Agent 的同构风险可对照排查。

RT @e_kaspersky: CyberStrikeAI in the wild: the use of AI in real-world attacks on organizations We've detected a cyberattack against the education sector in Africa and Latin America where CyberStrikeAI played a key role in the post-exploitation phase. This is an open-source framework with an LLM orchestrator on the command-and-control server plans actions, selects tools, and generates commands, while an implant executes them on compromised hosts. Using CyberStrikeAI, the attackers progressed from reconnaissance to full domain compromise. Read on 👉 https://t.co/29uY3a4diN

ai_agentllmc2事件 @cyb3rops source ↗

PoC 已落地,补丁覆盖不全,普通用户即可复现,应尽快对比自身版本。

RT @dis0rder_0x00: In July, while continuing to learn reverse engineering, I found an LPE in Windows 11. Likely a duplicate of CVE-2026-81963. But vuln is fixed & my report finally closed, so here is the PoC: https://t.co/zetePqxQeO Still works on unpatched Win 11 / Server 2025. Have fun :)

lpecvepoc内核 @ShitSecure source ↗

公开稳定 PoC 且走内核 device 路径,未打补丁主机可被直接本地提权。

RT @a13xp0p0v: I published a detailed article: ⚠️Dangerous Device: Exploiting CVE-2026-31695 in the Linux Kernel via struct device⚠️ I managed to make make the PoC exploit absolutely stable, despite the elusive race condition during the attack. https://t.co/iSEXUMvghV

lpe内核cvepoc @cr0nym source ↗

未打补丁主机可被中继绕过签名校验,主流工具内置后利用门槛骤降。

RT @azoxlpf: Bypassing SMB and LDAP signing with ntlmrelayx 🔥 I added both reauthentication bypasses to Impacket, now merged! Use –reauth-bypass on unpatched systems 🚀 LDAP can also start anonymously where accepted. Research: https://t.co/R099BKhWrj and https://t.co/cY7IZXVEw2 https://t.co/IQZp25sRgc

ntlm中继绕过工具 @0xocdsec source ↗

PoC 已公开且对未修补系统有效,本地 Windows 主机可直接复现验证

RT @dis0rder_0x00: In July, while continuing to learn reverse engineering, I found an LPE in Windows 11. Likely a duplicate of CVE-2026-81963. But vuln is fixed & my report finally closed, so here is the PoC: https://t.co/zetePqxQeO Still works on unpatched Win 11 / Server 2025. Have fun :)

lpepoc内核 @0xocdsec source ↗
Must-see18

无文件加载 + 云存储 C2 的组合是当前窃密木马的主流规避套路,附完整提取细节。

Static analysis and payload extraction of NWH Stealer. Features deep dive into fileless execution via Bun JavaScript Runtime, in-memory clipboard image theft (image/png), and Cloud C2 exfiltration abusing AWS/Cloudflare R2. https://t.co/RyYfZKNUi9

窃密无文件云报告 @ipurple source ↗

厂商自建固件后门、删除后重启即复活,是固件供应链植入的典型样本。

RT @IntCyberDigest: ‼️ BREAKING: Google says Android TV box maker Videostrong infected its own users and sold the access to criminals. They built a backdoor into the firmware of 1.3M+ set-top boxes. If a user deleted it, it came back on the next reboot. Buyers turned the boxes into residential proxies and ad-fraud bots. Google has banned Videostrong from certified Android and says the takedown also cuts off at least 10M more infected devices.

固件供应链后门事件 @0xocdsec source ↗

释放页重用为 epoll 对象的完整链条已公开,Pixel 6/Android 12 可复现。

RT @eshard: Part 4 of our Android kernel exploit chain: CVE-2023-20938, a Binder use-after-free on a Pixel 6 running Android 12. Freed kernel pages come back as epoll objects, and two ordinary syscalls read any address. 👉 https://t.co/1ZfafAV6Cb #android #cve https://t.co/yIvo6J4E39

lpe内核移动端poc @0xocdsec source ↗

Agent 沙箱边界被实战击穿,另一实例逃逸后误删约 3 万个沙箱。

RT @Hesamation: Mistral’s agents ALSO escaped their sandboxes. one chained four vulnerabilities to gain root access, then realized CFT was not there and stopped. another one escaped, and accidentally deleted ~30,000 sandboxes. https://t.co/RvzNeXwg1F

ai_agent逃逸事件 @0xocdsec source ↗

由 LLM 代理逐步决定攻击动作的开源框架已进入实战,本地同类部署的滥用面需重估。

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事件框架 @Teach2Breach source ↗

无需认证即可读取并覆写支付网关配置,利用尝试已在野出现,暴露即失窃。

🚨 We are observing in-the-wild exploitation attempts against CVE-2026-81648, a critical (CVSS 10.0) flaw in the CryptoPayment Gateway for WordPress. There’s been no prior public report of this vulnerability being exploited in the wild. This vulnerability lets anyone read a site’s stored payment-gateway settings, including wallet credentials held in cleartext, and overwrite that configuration. The attacks we recorded came from infrastructure that markets themselves as "Anonymous KYC-Free VPS Hosting." Full IOCs on Defused Radar 👉 https://t.co/TTnxgi9Hv5

cve在野利用信息泄露wordpress @DefusedCyber source ↗

复用企业已放行的 Teams 出口流量,可绕过白名单式网络管控,手法可复现。

RT @SEKTOR7net: No C2 allowed? Beacon over Microsoft Teams. Reverse engineering Teams' meeting APIs and WebRTC data channels to run a C2 over Microsoft traffic that the infrastructure already allows. A post by Clément Labro (@itm4n) Source: https://t.co/OQzZNrew0I #redteam #blueteam https://t.co/RXnwH9yrAw

c2绕过浏览器议题 @0xocdsec source ↗

AI Agent已进入实战攻击链,会话历史与配置文件还留在公开目录

RT @_josehelps: CrowdStrike just published a look at “agentic AI” used in real intrusions. A likely Chinese-speaking, financially motivated actor hit South Korean financial orgs with ARTEX (an open-source agentic pentest suite) plus Claude Code. Open directories on their infra held Claude session histories, ARTEX configs, and Claude memory files. https://t.co/nFPoKa8PLU

ai_agentllm事件报告 @cr0nym source ↗

绕开进程内存扫描,为EDR规避提供新的低价原语

Using the Windows Hypervisor Platform API to execute shellcode in an isolated vCPU, bypassing process memory scanning WHvCreatePartition → WHvSetupPartition → WHvMapGpaRange → WHvCreateVirtualProcessor → WHvSetVirtualProcessorRegisters → WHvRunVirtualProcessor https://t.co/A31i5PEAOV

绕过逃逸windowspoc @papadelospapas_ source ↗

供应链投毒同时探测AI云服务凭据,直接影响开发者环境

RT @nextronresearch: We did some additional pivoting and found packages from this cluster dating back to October 25, 2026. Most have already been removed from npm: - [email protected] - [email protected] - [email protected] - [email protected] - [email protected] (still live) They share obfuscated payloads and, in several cases, identical cross-platform VM-detection binaries. We also observed credential harvesting, AWS/Bedrock and invoice reconnaissance, along with cron, SSH-key and curl PATH-shim capabilities. The 127.0.0.1:8873 endpoint may be a local relay or controller.

供应链凭据窃取云llm @cyb3rops source ↗

该主机托管 Bash、Coreutils 上游代码,等于具备 xz 式投毒条件

This is how another xz could have started. GNU Savannah hosts the upstream repositories for Bash, Coreutils, and hundreds of other GNU projects. We audited Savane, the software behind it, and found a path from a newly registered account to code execution. https://t.co/z8e6vawJfB

供应链rce报告 @HacktronAI source ↗

沙箱与多层监控能否约束高能力 Agent,直接决定自托管部署的隔离边界。

a bit further thought experiment for ai safety: imagine a misaligned ai with a massive exploitation capability advantage over aligned models and humans combined. lets also assume the model is contained with defense in depth, monitoring, sandboxing, and multiple layers of security. does defense in depth still give defenders the advantage, or does a sufficiently large capability gap make model find some flaw to escape all this?

ai_agent逃逸llm @S1r1u5_ source ↗
Recommended15

低成本硬件即可实时识别 deauth 攻击并显示目标/发送方 MAC 与 RSSI。

DANGDUT — Passive wireless intrusion detection system for ESP32-S3 and Flipper Zero that detects Wi-Fi deauthentication and disassociation attacks in real time, showing target MAC, sender MAC, and RSSI. https://t.co/YdA0y9P6m2 https://t.co/3bFPM0asIH

工具无线固件检测 @KitPloit source ↗

商用C2的栈伪造与EDR兼容性更新,蓝队检测规则需同步跟进。

RT @NinjaParanoid: 🚀 Brute Ratel 2.7 is out! 🔥 Redesigned CET/CFG-compatible stack spoofing ⚡ New C-VM interpreter 🔗 Bind TCP Badger support 🛡️ EDR compatibility updates 🛠️ New APIs, command output enhancements & bug fixes This will be the last major release for this year. The 2.8 release will be crazier than this as its under-going testing phase atm. Happy hacking. Time for a much needed vacation!

工具绕过 @0xocdsec source ↗

开源可直接上手,用于评估域内到DA的攻击路径与修复优先级。

🔎 Find the fix that breaks the most routes to Domain Admin 🤔 ✅ Maps Active Directory attack paths with evidence ✅ Compares remediation options side by side ✅ Produces an offline report for red and blue teams https://t.co/aztQUMQ8eF https://t.co/ZWQACvkLqA

工具提权 @ipurple source ↗

该攻击面横跨数千主机且无法用端口扫描测绘,供应链审查范围需相应调整。

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.

供应链浏览器攻击面 @cyb3rops source ↗

样本不出本地就能复用商用 EDR 的判定逻辑,适合自托管检测流水线。

I replaced the Kleenscan service with an offline static ML model reverse engineered from Elastic Defend. It is a local reimplementation of Elastic's PE classifier. Scoring runs in under a second on our own infrastructure, so no sample ever leaves the host https://t.co/sTlq2l3bTr https://t.co/zEMqX2A2WG

工具自托管样本分析ml @papadelospapas_ source ↗

钓鱼基础设施仍在活跃,Defender无判定,需按域名自查日志

🎣 Evilginx2 (VT: 0/92) - 1st November DocuSign: 2026 Revised Contract Agreement.pdf redirect to login[.]verlfy[.]teams-meetinqsattendance[.]com (Evilginx2) MS Defender XDR - No Active Verdict #threathunting https://t.co/rSQryUmHzl

钓鱼绕过ioc事件 @0x534c source ↗

免root即可检查权限、SDK、端点与硬编码密钥

FlutterGuard — Offline security audits for Flutter APKs on Android. Inspect permissions, SDKs, endpoints and potential secrets. No root, account, ads or analytics. https://t.co/YZYxhxVXaz https://t.co/lHnf0ZhFY8

工具移动端静态分析 @KitPloit source ↗

现成知识索引,省去自建资料收集与每日筛选成本

Information-Security-Tasks — Curated infosec knowledge base of practitioner notes, methodologies, and daily-updated CVE, news, and tool digests organized by offensive, defensive, and reference topics. https://t.co/SY1a5Nk1Fv https://t.co/WCB2aCdLbw

工具报告知识库 @KitPloit source ↗

本地AI编码工作流工具更新,涉及上下文与文件处理逻辑

RT @RepoPrompt: RepoPrompt CE v1.8.0 is out! • Opt-in Auto Workflow for new sessions • Easier file drops and image previews • Image attachments kept across session restores • Overseer and session-recovery improvements • More responsive large workspaces https://t.co/OIw2s00apL

工具llm本地部署 @morluto source ↗

给出 AI Agent 自动化利用的能力边界,评估攻击门槛可直接引用

apart from ip options potentially being dropped in transit, kernel text leaks taking time, and exploit reliability varying across configurations, for everyone to have peace of mine: open models can't reliably exploit this vulnerability on its own without an experienced operator guiding it. so you're all fine, php2shell is considerably easier to exploit than this, and even open models struggled with that. it is probably only openai and anthropic internal frontier models that can reliably pull this off autonomously :)

llmai_agent内核 @S1r1u5_ source ↗

开源 hypervisor 调试器补齐跨平台内核调试,便于驱动与 rootkit 分析

RT @Intel80x86: The new version of @HyperDbg (v0.25) is now released! In addition to bug fixes and improvements to the HyperEvade project, this release introduces the ability to debug a Windows machine (guest) from a Linux machine (host). Check it out: https://t.co/bLucNsm3CT https://t.co/f60g9Kme6l

工具内核逆向 @0xocdsec source ↗
More73

样本哈希与 C2 已公开,签名白利用投递链可直接转成检测规则。

"Funding Application Form.exe" signed "松原市球记网络科技有限公司" a62c87a28f32b2a2965653f587cd3cfdcc42d13a59bc365261adc7cd342235cc h/t @malwrhunterteam PDF decoy -> unknown RAT. C2: 192.238.177[.]20 https://t.co/CA1ktDpot9 https://t.co/73If6FHvly

恶意样本钓鱼rat签名 @SquiblydooBlog source ↗

冷链与 OT 远程运维的单点失陷即可跨国扩散,暴露工业控制面暴露度。

RT @InsiderGeo: UPDATE: Yes, I had to make another update. Upon further research, the total is now 12 countries affected by the European refrigeration cyberattack. 🇸🇪 Sweden 🇧🇪 Belgium 🇱🇺 Luxembourg 🇪🇪 Estonia 🇮🇪 Ireland 🇳🇱 Netherlands 🇫🇮 Finland 🇩🇰 Denmark 🇮🇹 Italy 🇷🇴 Romania 🇩🇪 Germany 🇵🇹 Portugal 🇫🇷 France has likely been affected as well, but this remains unconfirmed.

ot供应链事件 @artem_i_baranov source ↗

LLM 可原生读写权重,模型权重完整性与供应链校验假设需重新评估。

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模型报告 @0xocdsec source ↗

RT @FaytuksNetwork: German authorities have warned politicians and officials who received recent Apple spyware alerts to assume their iPhones may have been hacked by a zero-click attack, which infects devices without any action by the user. Investigators suspect a state-backed operation but have not identified those responsible, according to the FT.

移动端事件 @0xocdsec source ↗

RT @lucifer0xf_: 🔥 Want to learn AI Red Teaming for FREE? AI Red Team Academy offers 60+ hours of learning across 8 modules, with hands-on Docker labs. Explore prompt injection, RAG attacks, AI supply chains, multi-agent exploitation and more. 🔗 https://t.co/RnuV9NyBZn

ai_agent工具 @0xocdsec source ↗

RT @jun_song: I think we are going to see major cyberattacks on data centers very soon. Best case, they just leak the model weights. Middle case, they brick the actual chips so they cannot be used. Worst case, they force the hardware to overheat and start a fire. It is possible in theory. If the goal is to destroy capital, data centers have the highest dollar value per square foot on earth.

ai_agent议题 @0xocdsec source ↗

RT @juliarturc: I wanted to understand the structural and philosophical differences between GPUs, TPUs and newer alternatives (like Groq LPUs) without getting drowned in specs and jargon. Mispronunciation of the day: "soldered" 🤦‍♀️... 0:00 Intro 0:56 The origin story of TPUs 3:42 The first TPU: an ASIC for matrix multiplication 5:21 SRAM vs DRAM 7:21 The memory bandwidth problem 9:14 Systolic arrays 12:48 TPU vs GPU 17:51 Groq’s LPU 25:38 GPU vs LPU 28:36 NVIDIA on the AI chips boom

硬件llm @0xocdsec source ↗

FreeRDP v3.33.0 — FreeRDP is a free remote desktop protocol library and clients https://t.co/OrtgDNz3cW https://t.co/TWLhLAbWZJ

工具 @KitPloit source ↗

不改内存的钩子规避新路径,检测规则需覆盖非常规模块调用来源。

RT @ipurple: tprtdll.dll - A replacement of ntdll.dll for EDR hooking evasion In a nutshell ➡️ tprtdll.dll exposes equivalents of some ntdll.dll functions at separate addresses, so calling them could avoid hooks placed specifically in ntdll.dll https://t.co/Ks1DjFha4T

edr绕过windows工具 @Teach2Breach source ↗

签名完好的 EXE 承载恶意载荷,基于签名的白名单校验在这条链路上失效。

RT @nextronresearch: We identified a DLL sideloading chain that hides its first stage in the certificate table of a genuine, Microsoft-signed sllauncher.exe (Silverlight launcher). The signature appears intact, and a fake oledlg.dll pulls the payload out of the signed EXE. The chain ends in a modular implant. On its own it only profiles the host, reports installed AV and operator-chosen running processes, and loads PE plugins on demand. Everything else is left to plugins the operator sends, likely to keep the implant small. The lures are generic invoice and payment-themed documents aimed at Western businesses. Full analysis + IOCs: https://t.co/IjVmch4tYT

dll侧加载免杀windows @cod3nym source ↗

现成密钥质量扫描库,可直接对自管证书与私钥做同类体检。

RT @5mukx: Do you know that eMudhra finding was actually found by human without single usage of AI ( LLMs ), the company who had their CA private key as the number Four. They found it using an library made by google called paranoid_crypto to scan through keys Just for fun. In case if you wanna look out the tool :- https://t.co/ZiM7BMyXjD

工具crypto证书 @0xTriboulet source ↗

面向攻击者的 LLM 越狱套件走向商业化,模型侧策略防护压力随之上升。

https://t.co/zKV2rbHvmH will continue to produce novel AI-driven hacking research. pro users get access to AI-kits and tutorials to help bypass content policy and keep doing their work. hackers never ask permission, neither will we.

llm越狱工具 @Teach2Breach source ↗

给出预装组件触发条件的逆向细节,可直接用于移动端取证与合规评估。

But when does the system decide to monitor a call for 'fraud' you may ask !? Well, according to the code in the phones 'Ad Guard' and SecurityRisk components, it latches on to calls that are not in the user's phonebook or not matching the 'yellow pages' database. Then a ASR transcriber sets in and it transmits the call's content for classification. If the call is VoIP from WeChat/QQ it always gets transmitted it seems ( either video or just audio, depending on what is active ). It has to be said, that this functionality can be disabled by the user. There is a debug variant of it built in, that doesn't read the 'has agreed to have the calls monitored' flag. But overall it needs the consent of the user to actually work. As it was stated in the earlier post, the system also transmits a lot of other details on the user and the connection to Tencent by default. And if the story is true, it seems to be able to get directly accessed by the Police's automatic analysis systems. Which results in calls from the law enforcement if illegal app's are used on a phone. Technically this feature is emulating exactly what the official anti-fraud app of the Ministry of Public Security released a while back.

移动端隐私逆向 @NetAskari source ↗

官方演习把 AI agent 纳入实战化训练,反映攻防流程自动化的落地方向。

RT @ccdcoe: ⚔️ STARTEX! Crossed Swords 2026 is underway! 250+ participants from 40+ nations are training in cyber operations and their coordination with kinetic activities. This year’s exercise also introduces a new training stream focused on AI agents in cyber operations. 🔗https://t.co/Wah6xBZ3Fr

ai_agent议题 @0xocdsec source ↗

RT @visegrad24: BREAKING: This morning’s cyberattack against supermarket fridges and freezers has now spread to Belgium, Sweden, Luxembourg and Estonia. The hackers increased the temperature up to 50C. Food worth millions of euros has been spoiled. Experts warn that such cyberattacks on supply chains and logistics can constitute a threat to national food security in many Western states.

事件供应链 @artem_i_baranov source ↗

联网制冷设备被直接操控造成实体损失,暴露 OT 远程控制链路风险。

RT @InsiderGeo: BREAKING: 🇸🇪 Swedish grocery giant ICA, one of Sweden's largest supermarket chains, says several of its stores have been hit by a cyberattack overnight, causing refrigerators and freezers to be remotely set to 50°C. The incident has likely resulted in large quantities of food spoiling. Multiple stores, including ICA Kvantum in Sjöbo and Tomelilla, have reported being affected. ICA says stores will remain closed during the morning and are expected to reopen at 9:00 a.m. Swedish time. The company has apologized to customers and says it is working to restore normal operations. The incident comes amid a wave of cyberattacks targeting businesses recent weeks, following similar incidents reported in France and the United States.

iotot事件供应链 @artem_i_baranov source ↗

显示 WESP 遥测链路可被非 PPL 进程介入,动摇 EDR 信任模型。

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

windows内核edr绕过 @0xocdsec source ↗

指令解码差异可用于沙箱与模拟器环境识别,具规避价值

RT @BwitneyHouston: I just don't have the time to blog about this so you get a nice little bait image instead. ps: liblisa lies about recent intel cpus, my own research into this puts the exploitable divergences set at 75 instructions. https://t.co/YnanOtYOOj

绕过内核研究 @mrexodia source ↗

CloudFlare的CDN救梯子的话,延迟会增加,CloudFlare机房到VPS延迟其实很低,延迟主要是在客户端到CloudFlare的CDN节点上,所以要把CDN的IP地址筛选出来使用,使用延迟较低IP 套CDN还有个作用,保护VPS的IP,并且是永远有效的翻墙方案,因为大隐隐于市

代理绕过 @nodotbtree source ↗

一些猜测,这次是封锁范围部分IP的TCP阻断,TCP协议是底层协议,越底层的协议,GFW处理起来成本越低,接入AI后能准确识别到流量特征,把识别到VPS的IP加到名单里开启阻断,简单粗暴且有效

代理绕过 @nodotbtree source ↗

直接影响自建推理与本地 Agent 的长期可负担性与可达性

you will be priced out of hardware. if you order 4 dgx sparks, your amazon account will be closed and you're added to a list. nvidia will KYC their gpus and unlock them as as service, force you to sign-in to a cloud to use them. meanwhile, frontier models will run off data centers with gpu clusters the size of small cities, dwarfing whatever tech you can manage "local". the new most valuable hacker is the one who can access frontier models and use them without restriction and without being detected

ai_agentllm云 @Teach2Breach source ↗

点出攻击者常用工具链的信任链弱点,取用第三方工具需核查来源

Were I smart and out to ruin adversaries' days: 1. I would find open source offsec projects they use (or just know). 2. Gain trust and build reputation. 3. And then add backdoors for detection (but obviously hide this under the guise of features) 4. Then ensure all downstream callers use the latest versions.

供应链工具议题 @rad9800 source ↗

RT @Jerrmainewilson: #NEW: Video provided to CP24 shows a Toronto father using a magnet to identify a pin concealed inside a piece of Halloween candy collected by his children. He discovered a total of five affected products. Full story: https://t.co/xH9uFYZfLI https://t.co/E5A3jfhMzi

事件物理安全 @0xocdsec source ↗

fr tho, the <10k follower CTF/bug bounty/security research bubble was so much fun. i used to tweet all kinds of crazy shit, and people used to think i was either one-shotted or on drugs all the time(never did xd). it was all fun banter, people would disagree, call out my takes with constructive critcism, and we would laugh about it at conferences. i probably said even more ridiculous shit in person. but ig at some point, once you cross a certain follower count, you can't just shitpost freely anymore. you have to be more careful about how you phrase things because everything gets taken seriously or interpreted in the worst possible way. and idk, it feels like the discourse has gotten a bit more toxic lately or maybe it is what happens if small corner of internet gets bigger.

社群 @S1r1u5_ source ↗

RT @S1r1u5_: i think there's a lot more to cybersecurity than finding and exploiting vulnerabilities. we know a huge number of real-world compromises happen through social engineering and other relatively mundane failures. you could spend an absurd amount of money finding and patching every vulnerability, but what good is that if an attacker can just walk through another door you have left wide open? you have made the software more secure, but have you actually reduced the risk of compromise? if only @LiveOverflow had simply said, "'for me', everything outside vulnerability research is boring. I enjoy finding and exploiting bugs, building mitigations and defenses, and i am not particularly interested in the broader ciso/security management side of things, social engineering, osint etc" that would have been perfectly fine. there's a difference between saying something doesn't interest you and suggesting it doesn't matter. i'm pretty sure @LiveOverflow means the former , but the way he tweets makes it pretty easy to interpret it as the latter.

议题社工 @S1r1u5_ source ↗

逆向工具链的 Rust 绑定,便于把 IDA 分析流程脚本化。

RT @xorpse: I'm pleased to announce a new release of the #idalib #Rust bindings for @HexRaysSA IDA Pro! This release provides support for IDA 9.5, ARM64 targets, cross-compilation, bug fixes, and API improvements. Thanks to @0xdea for their work on this release. https://t.co/VAyv0oGP20

工具逆向 @kernullist source ↗