⚡Megatron LLM Hub
All digests
Twitter

推特安全流

Top7

fetch() 默认路径可达,只需对端返回 br 编码,修复版本已发布。

RT @julianor: While everyone’s patching PHP today, don't sleep on Node.js CVE-2026-111497 OOB write in the bundled Brotli decoder. Reachable via fetch(): a malicious server just replies Content-Encoding: br. Fixed in 26.12.1 / 24.22.1 / 22.23.4. Update.

rcecvenodejs @0xocdsec source ↗

影响 WordPress/Drupal/Joomla 等主流站点,预认证即触发,无需应用层逻辑缺陷。

RT @sagitz_: Today we disclose PHP2Shell - remotely exploitable memory corruptions in PHP WordPress pre-auth RCE: ✅ Drupal pre-auth RCE: ✅ Joomla pre-auth RCE: ✅ Remote memory disclosure of any PHP application: ✅ Read more ⬇️🧵 https://t.co/wVfy4kdHhz

rcecvephp @artem_i_baranov source ↗

官方机构给出停服级建议,说明在野利用风险与补丁紧迫性。

RT @NSACyber: Effective Immediately, all organizations using PHP for their systems are advised to be prepared for an update addressing two critical vulnerabilities. PHP is widely used for web applications. These vulnerabilities often provide a direct path to the server's file system or database. If an immediate update is not feasible, it is recommend that organizations take the system offline until it can be fully mitigated. NSA will follow up shortly with more specific guidance.

cvercephp @0xocdsec source ↗

利用门槛低,补丁已出但存量部署多,Web 进程异常行为可直接落检测。

Two critical PHP RCEs are patched: CVE-2026-109618 and CVE-2026-109619, both CVSS 9.2. • Fixed in 8.5.12, 8.4.27, 8.3.36, and 8.2.35. You can still focus on detection coverage today. Security Labs writeup (coverage + guidance): https://t.co/aBB6g7gCZ5 Most PHP RCEs converge on the same outcome, whatever the bug: a web server process (php-fpm, php-cgi, apache2, httpd) doing something a request-serving daemon almost never does legitimately. • a shell to confirm access: id, whoami, uname -a • a second stage pulled in: curl http://... | bash • a reverse shell: bash -i >& /dev/tcp// • a webshell written into the web root The bug class is public now: out-of-memory corruption, and stale state reused by the same worker. The behavior is what you detect. We recently ran a different PHP RCE end-to-end (wp2shell): apache2 spawning dash to run id; whoami; hostname fired the exact rules that we suspect to catch php2shell. Payload Execution by Web Server kills the shell on the endpoint. The Suspicious/Unusual Command, Child, and File Creation via Web Server rules stack on top. (PHP did confirm partial php.ini mitigations. session.upload_progress.enabled=0 blocks the confirmed session paths only. Patch is still the fix.) Patch to 8.5.12, 8.4.27, 8.3.36, or 8.2.35. Verify that the stack is enabled now. Guidance + every rule link below.

rcecve检测规则 @elasticseclabs source ↗

自建 PHP 服务当天即可升级,官方称可能提前或遭实际利用

RT @ThePHPF: The PHP security team is giving advance notice: Upcoming security release Wed, Oct 21st 2026, 14:00UTC, two critical vulnerabilities, both rated CVSSv4 9.2. PHP users are encouraged to prepare to update. Earlier release if details become public or active exploit evidence.

cvercephp @virusbtn source ↗
Must-see41

AI Agent 驱动攻击链的实战证据,并关联 ConnectWise 供应链事件。

RT @MsftSecIntel: The China state-aligned threat actor Storm-0940 used the agentic AI penetration testing framework ARTEX to automate attacks targeting hundreds of organizations globally across a broad range of industries. Microsoft Threat Intelligence assesses that Storm-0940 was responsible for the recently disclosed ConnectWise supply chain compromise of a multitenant application. https://t.co/VXovj0jI90 Our investigation found that Storm-0940's use of ARTEX accelerated the speed of compromised application abuse to authenticate to target environments through applications that had already been granted access, allowing the threat actor to operate without requiring an interactive user sign-in. AI-accelerated follow-on activity included reconnaissance, application credential changes, and attempts to modify privileges that could enable expanded or persistent access. Read our blog for a detailed analysis and indicators of compromise (IOCs), along with Microsoft Defender detections, protection recommendations, and guidance on investigating service principal sign-ins and unexpected application credential or privilege changes.

ai_agent供应链事件报告 @cyb3rops source ↗

PHP 部署面极广,默认配置即暴露,需立即清点版本并监控利用流量。

RT @0xpvee: it took two PHP zero-days to get me to post my first tweet haha anyways, both are critical RCEs, meaning remote code execution with no login required. with PHP this widespread, defenders should be: inventorying PHP versions, watching for exploitation attempts and patching asap

rcecvephp @0xpwnie source ↗

即使自身不跑 PHP,共享宿主上的邻居中招也会连带受害,多数 VPS 商未修补。

RT @PaulosYibelo: there's a real risk that a single PHP server will be exploited to infiltrate a guest OS somewhere, & from there, CVE-2026-99999 could be exploited to breach the host OS, leading to all other guest servers. So even if you don't run php, if some random customer of your host does, you are still going to be affected. This is esp concerning knowing most vps/compute providers still aren't patched.

逃逸云phprce @h4x0r_dz source ↗

可由私钥直接预计算伪造任意站点证书,波及全球信任链。

In 2008, Debian disclosed that a change intended to silence Valgrind warnings had broken OpenSSL's random number generator. The change removed code that mixed entropy into its internal state, leaving the process ID as the only meaningful source of variability. Attackers could precompute the possible keys instead of breaking the cryptography. In 2026, eight of eMudhra's intermediate CA private keys turned out to be: 4, 56, 66, 115, 155, 156, 230, 237.

ca证书加密供应链 @PentesterLab source ↗

模型批量挖洞后,0day 首发归属与披露时效成新变量。

Everyone scrambling to claim PHP2Shell. Looks like a lot of collisions. It's something we've been talking about for a while now, new model drops, everyone scrambles to find 0days... model find 0day.. people rush for f4me. https://t.co/Fq24CjwlXJ

llm漏洞poc事件 @_xpn_ source ↗

影响覆盖任意 PHP 应用,公开 PoC 已在流传,需按版本尽快核对补丁。

RT @sagitz_: Today we disclose PHP2Shell - remotely exploitable memory corruptions in PHP WordPress pre-auth RCE: ✅ Drupal pre-auth RCE: ✅ Joomla pre-auth RCE: ✅ Remote memory disclosure of any PHP application: ✅ Read more ⬇️🧵 https://t.co/wVfy4kdHhz

rcephpcvepoc @0x64616e source ↗

给出实际利用态势,可用于判断补丁紧迫度而非只凭披露恐慌。

⚠️php2shell (CVE-2026-109618 and CVE-2026-109619) - we have so far not seen major exploit activity around either vulnerability. Exploit attempts seem to be limited to public POCs released in the past 24 hours, which based on our testing have predominantly been inoperative. Track php2shell exploitation on Defused 👉 https://t.co/iRDhHlDkep

cve事件php @DefusedCyber source ↗

触发点位于 OOM 与致命错误处理路径,与传统输入过滤防护无关。

RT @sagitz_: PHP2Shell (CVE-2026-109618 and CVE-2026-109619) are memory corruption issues in PHP’s internal error handling (out-of-memory, fatal errors). We've demonstrated that these can be exploited to full RCE or Heartbleed-style memory disclosure, regardless of application logic https://t.co/wkK6GS7As3

rcecve内存破坏 @artem_i_baranov source ↗

只需服务端返回 Content-Encoding: br 即可触发,补丁版本已明确列出。

RT @julianor: While everyone’s patching PHP today, don't sleep on Node.js CVE-2026-111497 OOB write in the bundled Brotli decoder. Reachable via fetch(): a malicious server just replies Content-Encoding: br. Fixed in 26.12.1 / 24.22.1 / 22.23.4. Update.

cverce内存破坏 @domchell source ↗

升级目标版本在此确认,运维可照版本号直接安排更新。

RT @php_net: PHP 8.4.27 released - 89 fixes in 30 components, including security fixes (CVE-2026-109618, CVE-2026-109619). Changelog + downloads: https://t.co/i0VF5tjG0x

cve补丁php @kernullist source ↗

给出 CVE-2026-109618/109619 的实际暴露面数据,可对照自查优先级。

RT @virustotal: Two critical PHP vulnerabilities (CVE-2026-109618 and CVE-2026-109619), a leak before the patch and millions of PHP servers on the Internet. We used VirusTotal Internet Scanning to measure who is exposed, what they run and how to cut the window: https://t.co/RgxlT3bCmr

cverce报告 @artem_i_baranov source ↗

展示 LLM 自动化漏洞挖掘的实际产出,并点名 OpenSSH 预认证面收窄点。

RT @chrisrohlf: Reducing the attack surface of deployed services and apps is a fundamental component of software that is secure by design. Earlier this year we tasked Claude with autonomously mapping and reducing the attack surface of 25 large and widely used OSS codebases. This blog post covers some of our methodology and a preview of the data we collected. This research resulted in a significant reduction in OpenSSH pre-authentication attack surface (now fixed in 10.6). And over 500 findings where reachability could be reduced and configurations hardened to ensure security controls weren’t silently disabled. https://t.co/8Iun038PNY

ai_agentllm报告 @S1r1u5_ source ↗

六个 TLS 中间 CA 私钥可恢复,能伪造任意域名的受信证书。

Let’s not forget about this certificate clusterfuck while everyone is busy with the PHP RCE news. Almost at the same time, another serious security incident became public. Eight intermediate CA certificates from eMudhra were found to use ridiculously weak private keys. We’re talking about P-384 keys where one private key was literally the number 4. FOUR! Six of these were TLS CAs, allowing anyone to recover their private keys and potentially issue trusted certificates for arbitrary domains. Major browsers have already blocked the affected CAs. But what about API clients, backend services, embedded devices, containers, and systems with outdated trust stores? Many don’t enforce Certificate Transparency or check revocation. Some could remain vulnerable for years. These weak keys have existed since April 2025. No confirmed malicious exploitation so far, but the implications are serious. Don’t overlook this one while focusing on PHP. Technical details and test endpoints: https://t.co/6NiX0zsfNm

catls证书事件 @cyb3rops source ↗

上游 8.5.4 未必存在可达路径,公开 PoC 的通用性存疑。

RT @UK_Daniel_Card: Testing the POCs in the lab with Claude, whilst drinking tea! (how 2026 of me :P /S cringe) and.... we have some interesting and useful intel..... exploitation of these CVEs doesn't appear trivial.... "Most likely why: the authors targeted a specific build (“Ubuntu php 8.5.4 profile-21”) with hard-coded struct offsets and a build-specific address delta. Vanilla upstream 8.5.4 either doesn’t carry that exact reachable path or the heap groom depends on that build’s allocator state. I’ll also be candid: this CVE pair is described in the PoCs themselves as leaked/pre-embargo/”presumed,” and I can’t independently verify it maps to a reproducible upstream bug." #PHP

rcepocphpllm @cyb3rops source ↗

可直接照抄的资产发现脚本,覆盖 CVE-2026-109618/109619 的排查

Two very useful posts on monitoring and hardening PHP services, plus an Nmap command to help find them in your environment: 1. Discovery with Nmap Common and alternative web ports; HTTP headers that may expose PHP versions: nmap -sT -sV --version-light -T4 --max-retries 1 --open -p 80,81,443,8000-8010,8080-8090,8181,8443,8880,8888,9000,9001,9080,9090,9443 --script http-headers --script-timeout 10s --script-args 'http-headers.path=/index.php,http-headers.useget=true' <target> Look for X-Powered-By: PHP/.... Fast inventory check; hidden PHP deployments can be missed, and version banners don’t confirm vulnerability. 2. Huntress: Monitoring and detection • Confirmed facts and remaining questions • Suspicious PHP child processes, shells and downloaders • Webshells and other post-exploitation activity • My process-watcher for Linux process visibility https://t.co/XLx8za50cZ 3. Vincent Schmalbach: Hardening and patching • Temporary nginx/Apache filters and upload-progress settings • WAF/origin bypass checks • Package and backport verification • PHP-FPM/worker restarts, container rebuilds • Side effects and checks for compromise https://t.co/30YfHmqIcR Temporary measures are stopgaps. Prepare to apply the official fixes as soon as available.

cve工具php @cyb3rops source ↗

官方修复尚在准备,暴露在公网的 PHP 实例已在无补丁窗口内。

🔥Two critical PHP RCE vulnerabilities, 𝗖𝗩𝗘-𝟮𝟬𝟮𝟲-𝟭𝟬𝟵𝟲𝟭𝟴 𝗮𝗻𝗱 𝗖𝗩𝗘-𝟮𝟬𝟮𝟲-𝟭𝟬𝟵𝟲𝟭𝟵 (𝗖𝗩𝗦𝗦 𝟵.𝟮), 𝗵𝗮𝘃𝗲 𝗯𝗲𝗲𝗻 𝗱𝗶𝘀𝗰𝗹𝗼𝘀𝗲𝗱 𝗮𝗵𝗲𝗮𝗱 𝗼𝗳 𝗮𝘃𝗮𝗶𝗹𝗮𝗯𝗹𝗲 𝗽𝗮𝘁𝗰𝗵𝗲𝘀. Emergency PHP security updates are being prepared, and affected versions are still being confirmed. ​ 𝗙𝗼𝗿 𝗻𝗼𝘄, 𝘁𝗵𝗲 𝗽𝗿𝗶𝗼𝗿𝗶𝘁𝘆 𝗶𝘀 𝗰𝗹𝗲𝗮𝗿: 𝗶𝗱𝗲𝗻𝘁𝗶𝗳𝘆 𝗲𝘃𝗲𝗿𝘆 𝗶𝗻𝘁𝗲𝗿𝗻𝗲𝘁-𝗳𝗮𝗰𝗶𝗻𝗴 𝗣𝗛𝗣 𝗶𝗻𝘀𝘁𝗮𝗻𝗰𝗲, 𝗮𝘀𝘀𝗲𝘀𝘀 𝗲𝘅𝗽𝗼𝘀𝘂𝗿𝗲, 𝗮𝗻𝗱 𝗯𝗲 𝗿𝗲𝗮𝗱𝘆 𝘁𝗼 𝗽𝗮𝘁𝗰𝗵 𝗶𝗺𝗺𝗲𝗱𝗶𝗮𝘁𝗲𝗹𝘆 𝘄𝗵𝗲𝗻 𝗳𝗶𝘅𝗲𝘀 𝗯𝗲𝗰𝗼𝗺𝗲 𝗮𝘃𝗮𝗶𝗹𝗮𝗯𝗹𝗲. 𝗗𝗼𝗻’𝘁 𝘄𝗮𝗶𝘁 𝗳𝗼𝗿 𝗮 “𝗽𝗵𝗽𝟮𝘀𝗵𝗲𝗹𝗹” 𝗲𝘅𝗽𝗹𝗼𝗶𝘁𝗮𝘁𝗶𝗼𝗻 𝗰𝗮𝗺𝗽𝗮𝗶𝗴𝗻 𝘁𝗼 𝘀𝘁𝗮𝗿𝘁 𝗯𝗲𝗳𝗼𝗿𝗲 𝘁𝗮𝗸𝗶𝗻𝗴 𝗮𝗰𝘁𝗶𝗼𝗻. https://t.co/JWWesPE0fU ​ #php2shell #vulnerabilitymanagement

rcecvephp补丁 @0x534c source ↗

改一行 php.ini 即可临时断链,补丁还需数小时到数天才能铺开。

RT @Kostastsale: A moron has released a POC early for php2shell so here is what you need to do: If you’re on version prior to 8.2: Edit php.ini: session.upload_progress.enabled = Off Then reload php-fpm and restart apache If you’re on later version: Update (it’ll be available in a few hours) In any case, the mitigation for the php.ini should work either way. Not sure if it’ll cover both vulns though so keep an eye on that. I’ll be reporting on follow up activity on windows as a target. Good luck!

rcephppoc @nas_bench source ↗

客户端仍信任该 CA 时,中间人可冒充银行等站点,影响面覆盖所有用户。

A stolen website key puts one site at risk. eMudhra’s weak CA keys could let attackers forge HTTPS certificates for any domain. If they intercept your connection and your client still trusts those CAs, they could impersonate your bank and read or modify traffic.

catls中间人 @OSdev_ source ↗

上传路径默认可达,利用思路已公开而官方暂无缓解措施,PHP 站点需立即评估。

RT @rina4_tot: More details about CVE-2026-109618 and CVE-2026-109619, the PHP RCE. Upload progress can serialize before basic RINT clears BG(serialize) There is currently no possible mitigation. See the screenshot below to understand how I got the UAF 😼 https://t.co/wdUjCP2hRa

phprceuafcve @0xocdsec source ↗

AI 辅助漏洞挖掘已规模化产出真实漏洞,攻防节奏与补丁压力将同步上升。

RT @oct0xor: Dr. Zhiniang Peng (@edwardzpeng) reveals how he discovered 200+ confirmed vulnerabilities in Microsoft Windows with the help of AI. The game has changed completely. 🤯 https://t.co/zN5xbKvpzj

llm漏洞挖掘windows议题 @0xocdsec source ↗

AI 直接产出二进制使能力约束更难施加,对 AI 攻防边界的判断需重估。

刚开始接触生物学的时候,我经常想象原始海洋里的有机分子构成核糖核酸,然后诞生了最早的单细胞生物,单细胞生物又聚在一起变成黏糊糊的原始多细胞生物,最终一步一步演化出各种生灵。我也经常想,如果能在实验室里加速重现这一过程,完整观察整个生物演化史就好了。 刚开始接触计算机的时候,我想过一个问题:程序是从哪里来的?后来我知道了编译器,程序从编译器来的。编译器自然也是从其它编译器来的。那么最早的编译器是从哪里来的?后来我逐渐搞清楚了这个问题的答案。但偶尔还是会想一些相关的问题,比如:如果现在全世界的编译器都忽然消失,人类要怎么重建整套体系呢? 最近几年研究 AI 安全问题的时候,我经常想:如果只给予 AI 完成特定工作所需的工具,那么 AI 的能力是不是就会被限制住?是不是就可以阻止 AI 获得强大的网络入侵能力? 2026 年 2 月,我看到马斯克预测到年底 AI 就可以绕过编译过程直接创建二进制文件。当时我做了一些实验,发现虽然 AI 还不能创建完整可运行的程序,但已经可以生成正确的机器码片段了。到了 9 月,我再次实验,发现此时 AI 已经可以用直接创建二进制文件的形式输出不太复杂的程序了。 但用直接输出二进制文件的方式去写大型程序显然不是好的选择,即使能写出来,也难以维护。那么,是不是可以让 AI 重演现代编译器的诞生过程?就像重演从核糖核酸到哺乳动物的生物演化史? 下面这个视频里超级马里奥游戏就是从核糖核酸开始构建的。 AI 先是通过直接输出二进制的方式构建了一个 3584 字节的机器码装配器 A0。然后用 A0 生产了一个更类似汇编器的工具 A1。又用 AI 生产了最基础的 C 编译器 C0。再一步一步迭代,最终演化出了一个完整支持 C89 标准并部分支持 C99 标准的编译器。有了这个编译器,就可以写出超级马里奥游戏,或者是任何网络攻击程序。 对 AI 说:“要有光”,然后整个世界就一步一步诞生了。所以我把这个项目命名为 BitKindle。感兴趣可以测试一下:https://t.co/GN9onM4vTq

llm二进制ai_agent议题 @tombkeeper source ↗

PHP 部署面极广,若默认安装即受影响,暴露规模远超单一应用。

RT @h4x0r_dz: if this affects the default installation of PHP We are fucked up CVE-2026-109618 CVE-2026-109619 PHP RCEs with a CVSS score of 10.0 ☠️ https://t.co/kr47MzTvIj https://t.co/NDq0mppHoT

rcecvephp @0xocdsec source ↗
Recommended15

跨请求内存破坏在 Apache prefork 下继承父进程映射,利用门槛低于预期。

RT @ikotas00: 話題のPHP2Shellですが、弊社ではパッチ公開前にExploitの構築に成功しておりました。複数リクエストにまたがるメモリ破壊であり、Apache等のprefork構成では、親プロセスでマップされたアドレスが子プロセスへ継承されるため、ASLRだけでは十分な緩和になりません。早急なアップデートを推奨します。 https://t.co/JzNd3orxKQ

php内存破坏绕过poc @0xocdsec source ↗

把 PHP 内存破坏漏洞的影响面落到自托管边界设备,升级优先级明确。

RT @cantcomputer: Friendly reminder with today’s PHP2Shell disclosure: pfSense is heavily PHP-based. Its entire WebGUI and much of its management logic run on PHP.

php固件rce边界设备 @0xocdsec source ↗

防守方检测规则可被离线提取,绕过门槛下降,检测逻辑应假设已泄露。

RT @Kostastsale: I used Opus 5.5 on max and GPT-6 Astra for 5 days straight and I reversed engineered Elastic Security and extracted all the rules. I made them available here: https://t.co/fGAoNleFEW You’re welcome.

llmedr工具绕过 @_EthicalChaos_ source ↗

本地 Agent 与桌面环境深度集成,值得评估其权限与调用边界。

RT @typecraft_dev: oma-cast has already received dozens of PRs! My favorite by @tcballard was just merged. Now you can ask your default agent configured in omarchy when you start typing in the search box. This is a fun project for @omarchy ! check it out here: https://t.co/00dgb2rEzv https://t.co/sdAtx0qXXX

ai_agent工具linux @morluto source ↗

前置 WAF 场景可立即降风险,同时明确仍需升级到已修补的 PHP 版本。

RT @BraydenWilmoth: One-click mitigation for php2shell vulnerability: If you have Cloudflare in front of your PHP website, please login to the dashboard and click this button to make sure you are protected. https://t.co/KywdLV8TK7 This reduces your risk but doesn't fix the vulnerabilities in PHP. We recommend updating PHP to a patched version.

rce云缓解 @cyb3rops source ↗

可直接搭靶场验证检测规则是否生效、补丁是否覆盖到位。

RT @Dinosn: Releasing my lab as well PHP RCE https://t.co/VcWBy1p68A , use it as a guidance and further reference for work upon. Patches are available at this moment PATCH! https://t.co/SdPh2WoZuL

pocrce工具 @artem_i_baranov source ↗

补丁窗口期内以容器隔离加进程监控兜底的现成检测工具。

RT @cyb3rops: Since I heard about this #PHP #RCE last week, I've been preparing a tool for everyone who doesn't know what to do right now: process-watcher shows you every program your PHP processes start, with full command lines, so you can spot suspicious activity Until a patch is out, isolation (containers) and tight monitoring are your best bet to limit the damage. And I'm sure the LLMs already go "brrrr" and people are trying to reproduce the vulnerability. From what I've heard, exploitation will probably show up as suspicious sub process spawns. That's exactly what this watches, using eBPF: one static binary for Linux (amd64, arm64), no dependencies, running in a few minutes. It can also kill the ones you name (dry run first), but it's a detection layer, not a patch, and it can't see anything that happens inside PHP without starting a process. Quick start: https://t.co/7nluFV3Q25 Project: https://t.co/45agyRKtnS

工具rce监控 @artem_i_baranov source ↗

自托管多 agent 攻击编排的参考实现,可直接在隔离环境复现实验。

ARTEX — Multi-agent LLM system for autonomous penetration testing in isolated labs: planner/worker agents share asset and exploration graphs, with MITM traffic recording and human-in-the-loop approval. https://t.co/t4D78apGWX https://t.co/0dPSJT0slL

ai_agent工具llm @KitPloit source ↗

官方自评 9.2,公网 PHP 资产需预留应急升级窗口。

RT @ThePHPF: The PHP security team is giving advance notice: Upcoming security release Wed, Oct 21st 2026, 14:00UTC, two critical vulnerabilities, both rated CVSSv4 9.2. PHP users are encouraged to prepare to update. Earlier release if details become public or active exploit evidence.

cverce @artem_i_baranov source ↗
More155

分层载荷链条细节具体,检出率低,可据其补充 LNK 与 certutil 检测规则。

RT @nextronresearch: #Kimsuky (APT-C-55) is delivering persistent backdoors through multilingual LNK phishing lures After reading the 360 Threat Intelligence Center's report on APT-C-55's multi-path delivery campaign, we found a cluster of nine related malicious LNK files submitted between September and October 2026, with detection ratios around 10-30 out of 60+ engines. The LNK files execute a layered attack chain: base64-encoded batch scripts decoded via a renamed copy of certutil.exe, staging in C:\Users\Public, nircmd disguised as ntoskrnl.exe, and scheduled tasks running every 9 minutes for persistence. A parallel variant uses PowerShell with double hex-encoded payloads, and a third path poisons the open-source pybroker Python library as a supply chain attack. Analysis & IOCs: https://t.co/QbhNDXUZvD

钓鱼lnkapt报告 @cyb3rops source ↗

RT @HackingLZ: Something that gets lost in all these conversations is that there were years when a major 0day would drop and technical details were almost impossible to come by. Multiple people here would stay up for days reverse engineering the issue and stitching together a working PoC, not just to prove exploitation was possible, but to help defenders understand the actual risk, develop IoCs, build detections, and clean up compromised environments.

0day报告 @0xTriboulet source ↗

签名不再是可信信号,白名单需按发布者证书而非签名状态校验。

@malwrhunterteam shared two MSI installers with this signer "SOPHISTICATED SOFTWARES (PTY) LTD" The first is a rust remote shell https://t.co/tUoOUV8j7N The second is a tool which calls itself CrocoRAT: https://t.co/YIKUVrSINt We've seen CrocoRAT a few times 🧵

供应链windowsrat钓鱼 @SquiblydooBlog source ↗

在野利用早于公开披露,日志回溯的排查窗口需显著拉长。

Someone I know very well ran scans through old boxes and found traces of this PHP exploit going all the way back to 2022 But many people on this website prefer to believe that sophisticated nation states don’t have actual genius hackers and nothing is real until someone is blasting it out publicly for nerdpoints

php事件rce @C2IRIS source ↗

漏洞研究产能与披露节奏被改变,防守方修复窗口被压缩。

gotta give it to openai and their sane daybreak program heif heist by harsh kvmescape by paulos php2shell by lcfr gitlab rces by joaxcar chrome full chain by moyix wordpress rce by hashkitten all largely thanks to openai daybreak giving researchers access to gpt 5.6 sol. anthropic took the opposite approach, gatekeeping access, there is a lesson to be learned here

llmai_agentrce @S1r1u5_ source ↗

AD/Entra 攻击链资料集中成档,便于自查与复现验证。

RT @cr3ghost: If you're into Active Directory, Entra ID or identity security, bookmark this. Researcher @_dirkjan has built one of the best FREE identity security research archives online. The rabbit hole covers: - Active Directory internals - Kerberos - NTLM relaying - Forest trusts + SID filtering - AD CS - mitm6 + krbrelayx - Zerologon - BloodHound - Entra ID internals - Primary Refresh Tokens - Windows Hello for Business - Conditional Access bypasses - OAuth + token abuse - Cloud Kerberos Trust - AD -> Entra lateral movement - Entra -> AD lateral movement - Temporary Access Pass abuse - Federated credentials - Managed Identities - Intune certificate abuse - ROADtools + ROADrecon + roadtx - Actor Tokens Some absolute classics: - Relaying Kerberos over DNS - NTLM relay to AD CS - Active Directory forest trust bypasses - Domain Admin from Entra via Cloud Kerberos Trust - Phishing Primary Refresh Tokens - Windows Hello key abuse - Conditional Access bypasses - Global Admin in every Entra tenant via Actor Tokens There are also years of FREE conference slides, demos and presentations. Blog: https://t.co/8qEBsLuiPy Talks + slides: https://t.co/5neZc1JAnQ Huge credit to @_dirkjan / Dirk-jan Mollema for publishing so much deep identity research openly. #ActiveDirectory #EntraID #RedTeam

ad身份工具报告 @cr3ghost source ↗

AI 驱动的 N-day 收集正在改写漏洞披露与赏金生态。

HackerScam is telling security researchers to send them php2shell valid POC lol They just want to collect POCs for their ai After that they gonna close your report as n/a Because 100% of the bug bounty programs do not accept new N-day https://t.co/ryBW3iWzsW

ai_agentpoc议题漏洞赏金 @h4x0r_dz source ↗

自托管多 agent 编排的风险或取决于系统间交互而非单模型。

Watching agents cohabitate is fascinating. This trace in particular reminds me of this paper https://t.co/alsAtq8J4Y My bet is that understanding the systems agents form together will become a bigger constraint on autonomy alongside humans than individual model capability.

ai_agentllm议题报告 @matterpreter source ↗

多版本变体并存,追溯与查杀的样本线索更具体。

@fr0s7_ @malwrhunterteam @smica @skocherhan @aaqeel87 Yeah, it’s being distributed a lot across Europe and US now. There are several versions floating around and some versions go under the guise of “Winos”, and that was leaked to GitHub in the past.

恶意软件威胁情报windows @d4rksystem source ↗

侧面反映逆向与 AI 交叉社区活跃度。

i originally thought people were afking but then I stopped by and realized that was not the case met some really cool community members like @Akhnatouch @ForcelineHD &amp; had great discussions on creating, reverse engineering and even ai research

社区逆向 @morluto source ↗

剩余 WSP 与 PrinterBug 无法中继到 LDAP,强制认证攻击面收窄。

RT @RedTeamPT: 🦛🦛🦛Microsoft patched PetitPotam this October: EFSR now only resolves local paths. This means WebDAV to LDAP Coercion is fixed! Remaining methods for clients are WSP (coercing SMB) and PrinterBug (coercing RPC), but both cannot be relayed to LDAP 💀💀💀

提权绕过ad @0x64616e source ↗

补丁二进制提前可 diff,攻击者可先于多数用户定位漏洞点。

RT @mikko: "On October 21st at 07:41 UTC, more than six hours before the coordinated release, Chainguard unilaterally broke the embargo and published built images containing the fix. Like other distributors, Chainguard had received the patch in advance so its users could update the moment the vulnerabilities were disclosed. Any attacker could now study the fix in binary form while every other PHP user was still waiting for the release. This disregarded weeks of work by many parties toward a coordinated disclosure." https://t.co/hszd7sh0NI

供应链php事件 @0xocdsec source ↗

可直接封禁的扫描源,显示披露前已有在野试探。

RT @craiu: Currently seeing 196.189.236[.]67 probing for the PHP vuns, suggest to block them on gateways, been hitting honeypots with a bunch of stuff since at least Oct 2.

ioc扫描php @0xocdsec source ↗

提供低成本部署侧缓解思路,可对照现有 PHP 运行方式核查。

RT @dcolascione: PHP exploit appears to need ASLR leak. I run my PHP via old-school cgi-bin, one /usr/bin/php process per request, which means no cross-request ASLR info leak. I guess I'm not vulnerable. My setup might be the the most secure (if not exactly fastest) way to run PHP.

php缓解部署 @0xocdsec source ↗

RT @calif_io: Let’s set the record straight: @lcfr_hax was the first to discover php2shell and develop a working exploit. See his August 5 tweet: https://t.co/I88EQFrUHa His work convinced the PHP Foundation to invest in fixing this entire bug class. He then worked around the clock to review, test, and retest the patches.

事件php @artem_i_baranov source ↗

is it reasonable to update our estimates of cyber risk downward if we repeatedly see catastrophic vulnerabilities like php2shell, kvmescapes, browser full chains, and compromised tls certificates without catastrophic outcomes? or is cyber risk heavy-tailed, where a rare event could be unimaginably bad? but it seems that never happened, so shouldn't we still update downward?

风险漏洞 @S1r1u5_ source ↗

RT @IntCyberDigest: ‼️ WhiteDate, the neo-Nazi dating site German Pink Power Ranger hacker "Martha Root" wiped on stage, now forwards to GoebbelSuite, a Google parody built on the AfD leak. https://t.co/NzPUBFYrFV https://t.co/rhB3NDRXMQ

事件defacement @0xocdsec source ↗

展示 LLM 参与真实漏洞挖掘并提前协调披露的完整流程案例。

RT @OpenAI: OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites. Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors. Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact. This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare. https://t.co/TW3jrSbB3f

llm漏洞报告 @artem_i_baranov source ↗

提示即使自身不跑 PHP,共享宿主的邻居被攻破也可能牵连。

RT @PaulosYibelo: there's a real risk that a single PHP server will be exploited to infiltrate a guest OS somewhere, & from there, CVE-2026-99999 could be exploited to breach the host OS, leading to all other guest servers. So even if you don't run php, if some random customer of your host does, you are still going to be affected. This is esp concerning knowing most vps/compute providers still aren't patched.

逃逸云rce @0xocdsec source ↗

规则以开源文件形式给出,便于移植进自有 EDR 或 SIEM。

Rules it maps to: • Payload Execution by Web Server: https://t.co/KEYOuntUqu • Payload Downloaded via Curl or Wget by Web Server: https://t.co/kJq3QqwGXf • Suspicious Command Execution via Web Server: https://t.co/SWmo1iTOqv • Unusual Command Execution via Web Server: https://t.co/hojx0VYqMb • File Creation via Web Server (+ polyglot / double-extension): https://t.co/s1ZJgS5b5w wp2shell walkthrough: https://t.co/T4ulNVyQoP

检测规则rce工具 @elasticseclabs source ↗

披露影响面评估与厂商协同打补丁的经过,可判断严重度。

RT @itszn13: php2shell is the most consequential vulnerability I have ever helped discover. I am proud of how our team worked closely in coordination with the PHP Foundation to provide support and to harden the patch in preparation for this release.

Please read more about our approach: https://t.co/dBlfRky7IL

rcecve报告 @moyix source ↗

RT @justdionysus: Fancy human generated comics, a good story, and an annoyingly pervasive root cause. Yes, it’s exploitable. Yes, it’s exploitable under ASLR. Yes, disclosure was eventful. @lcfr_hax hacked the planet and then tried to patch it

rce议题 @lcfr_hax source ↗

补丁本身已成为漏洞研究上下文,旧补丁的防护假设需重新验证。

RT @orenyomtov: The rate of patch bypasses has skyrocketed (8x) this year. 🚀 The good news is that the timing of the increase happens to correlate with Anthropic and OpenAI's security product releases so the increase might be defender-lead. The bad news are that vulnerabilities that have been reported to be exploited in the wild were bypassed 4x more often, so attackers definitely also got the memo that security patches are highly valuable context for vulnerability research. For defenders, it's getting important to spend considerable resources on robust and comprehensive patching to stay a step ahead of the attackers. More stats, like in which linux kernel areas 10%(!) of vulns are directly caused by patches of other vulns, are available in the full report below

报告绕过llm @cyb3rops source ↗

对 LLM 自动挖洞能力的现实校准,避免高估本地模型产出。

there's still a considerable gap between open and frontier models, and an even bigger gap between frontier models alone and frontier models in the hands of skilled researchers. a huge number of people threw tokens at the bug and produced mostly slop PoCs. the benchmark saturation you see online often doesn't translate into frontier level offensive capability. sure, deepseek flash and kimi k3 can probably develop exploits for a good number of vulnerabilities, but the capability gap is still substantial. they mostly failed for this php bug. prove me wrong. having a more capable model than your adversary is an advantage, and pairing it with a skilled researcher can make that advantage even greater for defenders. everyone involved in this research had access to the best cyber models. i personally think we should preserve that capability gap rather than dismiss its defensive value without evidence.

llmai_agentpoc @S1r1u5_ source ↗

补丁落地前的临时缓解配置,仅覆盖部分漏洞,不能单独依赖。

RT @Kostastsale: A moron has released a POC early for php2shell so here is what you need to do: If you’re on version prior to 8.2: Edit php.ini: session.upload_progress.enabled = Off Then reload php-fpm and restart apache If you’re on later version: Update (it’ll be available in a few hours) In any case, the mitigation for the php.ini should work either way. Not sure if it’ll cover both vulns though so keep an eye on that. I’ll be reporting on follow up activity on windows as a target. Good luck!

rce加固 @artem_i_baranov source ↗

运行时同属暴露面,PHP 之外还需同步升级 Node 各 LTS 线。

RT @itsMikeKipruto: Not just PHP. Node.js has also announced security releases for October 21, with HIGH-severity vulnerabilities affecting its 26.x, 24.x and 22.x release lines. If you run either stack in production, keep an eye on official advisories and apply patches promptly. Security maintenance doesn't stop at your application code. Your runtime matters too. Node.js advisory: https://t.co/rmgw4sOnqU

cve运行时 @artem_i_baranov source ↗

快速摸清资产中 PHP 版本分布,决定补丁与下线顺序。

RT @cyb3rops: PHP Patch, take it offline, or put it behind a WAF and hope it filters the exploit requests. I wouldn’t count on a configuration-only fix unless one is confirmed. For inventory, this Nmap command can reveal PHP versions exposed in HTTP headers: nmap -sT -sV --version-light -T4 --max-retries 1 --open -p 80,81,443,8000-8010,8080-8090,8181,8443,8880,8888,9000,9001,9080,9090,9443 --script http-headers --script-timeout 10s --script-args 'http-headers.path=/index.php,http-headers.useget=true' <target> Only scan systems you’re authorized to test. This finds advertised versions, not every PHP server, and doesn’t confirm vulnerability.

工具资产测绘 @artem_i_baranov source ↗

RT @attrc: Full details of a remotely exploitable vulnerability in the core engine of *all* PHP versions will be released tomorrow, October 21st. If your organization has internet facing PHP applications then you need to be ready to patch the minute the update is released or *strongly* consider taking the application offline until you can: https://t.co/hpdTOdvSsQ

rcecve @artem_i_baranov source ↗

AI 辅助研究得出的漏洞结论若不经独立验证,会引发全网误报。

Last night's slopfest surrounding the bogus PHP exploit and the subsequent panic wave is a perfect example of why good cybersecurity researchers are still needed, why LLMs are not infallible, and why critical thinking cannot be automated away. AI can accelerate research, but clanker confidence is not evidence and consensus built on unverified claims in the end is still just speculation. Next time we might not be so fortunate, so maybe as a community we need to tighten and review how we handle patching and assume leaks will happen and be prepared to counter that.

llmai_agent报告 @Laughing_Mantis source ↗

国家级机构已作响应,PHP 资产清点与补丁窗口需提前安排。

RT @ollieatnowhere: At the @NCSC we are asking all UK organisations and their supply chains across product and services to understand their use of PHP and prepare to patch... https://t.co/UPT31kqBL6

phpcve事件 @domchell source ↗

指纹伪装与 LLM 工具链结合,账号接管门槛进一步降低

RT @_CPResearch_: Our latest research exposes how AntiDetect browsers enable fraud, account takeovers, and coordinated influence operations at huge scale. Knowledge bases, 24/7 support and MCP servers for LLM integration, the ecosystem is evolving fast. Read more --&gt; https://t.co/g4vHtqwKx8

ai_agent反检测报告 @virusbtn source ↗

macOS 罕见变形样本,哈希与签名检测手段会失效

RT @objective_see: It’s not every day we encounter metamorphic ransomware targeting macOS! 👀 Our @patrickwardle recently presented at @macsysadmin &amp; @virusbtn on “Heartbreaker,” a new ransomware sample that rewrites &amp; recompiles itself every time it runs! Blog &amp; slides: https://t.co/ghYQTrAXFL

恶意软件macos报告 @virusbtn source ↗

披露了从 VTL0 内核写绕过凭据保护的潜在研究路径

Credential Guard has a shutdown mechanism inside LsaIso.exe. Signaling it causes the trustlet to gracefully unregister its Kerberos and CredMan RPC interfaces Currently researching: can a process with kernel write in VTL0 patch the object’s DACL and signal it from usermode?

windows提权研究 @Salsa12__ source ↗

RT @doomsxroll: @ipurple @_JohnHammond 100% agree. Skilled humans with domain specific knowledge still need to ask the right questions, and evaluate answers - AI is an agent not an actor and cant generate its own intent or motivation. AI isnt replacing anyone with domain specific knowledge. Learning is still valuable.

ai_agent观点 @ipurple source ↗

RT @GabrielGonzalez: Reverse Engineering Mask ROM Progressing a bit more on the #Silicon #RE side, I came across the structure used for manufacturing ROM (from the same paper as last time)... 🔗 https://t.co/JGC6UjRV6Z #HardwareSecurity #Infosec

硬件逆向 @cgomezz_23 source ↗

The real problem is there’s very little chance that it provides $1,000 in increased value to customers. The PS5 jailbreaks have given console insight into exactly the kind of features that customers care about the most. Unfortunately those are the same features that console has been moving away from for years now.

主机观点 @0xTriboulet source ↗

RT @ornith_: Aloha!🌺 Meet Ornith-1.5-122B: No Tradeoff Between Strong and Fast Stronger. Faster. Leaner. Ornith-1.5-122B delivers frontier-level capabilities with exceptional inference efficiency. With superior instruction following, stronger task decomposition, and deeper understanding of real-world workflows, it achieves state-of-the-art performance among open-source models of comparable size, while running on a single NVIDIA DGX card!! Open weights under the MIT license. Coding ✅SWE-bench Verified: 0.845 ✅TB2.1: 0.813 ✅CyberGym: 0.800 ✅NL2Repo: 0.512 Agentic ✅HLE: 0.446 ✅BrowseComp: 0.785 ✅MCP-Atlas: 0.746 ✅Toolathlon: 0.708 🤗Hugging Face: https://t.co/w1hD6rra1o We can’t wait to see what you build with Ornith-1.5-122B.

llm开源模型 @0xTriboulet source ↗

在抢修喧闹中提供反向判断,避免按幻觉 PoC 做错误处置。

i don't think there's a working public exploit. people seem to be just hallucinating with models, the bug is hard to exploit. also, it's wildly irresponsible to publish anything about the bug before the embargo lifts. even hinting at the technical details is dangerous. please don't do it

rcepocllm @S1r1u5_ source ↗

服务名与可执行路径不匹配是低成本高命中的排查点。

We had a Windows machine where one process kept coming back after every reboot. At first, it looked like normal Windows activity because the service name was: WindowsTelemetry That sounds believable enough. Windows has plenty of telemetry components, so if you are checking quickly, you might not even stop there. But I checked the full service configuration. The executable path was: C:\Users\kevin\AppData\Roaming\Microsoft\Telemetry\telemetry.exe That did not make sense. A service running as LocalSystem should not normally be launching an unsigned executable from a user’s AppData folder. I checked the file signature. NotSigned Then I checked when the service was installed. Event ID 7045 in the System log showed that WindowsTelemetry had been created the previous night around 11:08 PM. The service was configured to start automatically and run as LocalSystem. That meant the binary was not just persistent. It was running with very high privileges. I pulled the PID for the service and checked its network activity. The process had an established HTTPS connection to an external IP address. At that point, we isolated the machine. The interesting part for me was how normal the service looked from the name alone. WindowsTelemetry Microsoft\Telemetry telemetry.exe Everything was named to make you move past it. But once you looked at the path, signature and service account, the picture changed completely. This is why when I am reviewing Windows services, I do not only look for strange names. I check what they actually execute. I check where the file lives. I check who signed it. And I check which account the service runs under. A malicious service does not need a suspicious name. Sometimes it just needs a name boring enough that nobody bothers to open the properties.

windows持久化检测 @Officialwhyte22 source ↗

暴露 AI 生成漏洞利用在真实应急中的信息污染问题。

i think this evening-into-morning has been a fascinating storm of chaos and uncertainty with potential proof of concept exploits (generated by LLMs) being shared publicly and privately (then analyzed by LLMs) without clarity or understanding of what is actually "the" exploit or if it is "other" exploits (found by LLMs) all to hope for potential defenses (built by LLMs) we used to be so concerned with robots hallucinating at this point we might be the ones hallucinating ourselves

llmpoc事件 @_JohnHammond source ↗

攻击链与目标清单完整公开,可对照排查同类暴露面。

RT @tuitesteban: No defender caught this Russian-speaking operator. They went after #Ukraine’s state data-exchange system, and the working directory sat open for 17 hours... That is where everything was: the access log that proves execution, the scan output, the full target list The https://t.co/42EL2mKsZ6 research team rebuilt the whole campaign from it: a solo operator going after a state data-exchange system with a borrowed Log4Shell chain, and we only know its shape because they left the door open! Here is how we did it 👇 https://t.co/ZYQqsU45Ew

log4shell事件报告 @NetAskari source ↗

Agent 自恢复能力提升,同时意味着权限边界需要重新评估。

we're going to make rea a lot more robust in the upcoming v6.8 update ive been reading the posts + watching youtube videos @tonbistudio there will be a big focus on making it more comfortable for agents to use + adjustable so they can directly recover without human input

ai_agent工具 @morluto source ↗

it's interesting that eMudhra is a family-run business. not suggesting any relationship between how it's run and its security practices or innovation, but i can't help thinking about @captgouda24 "What's the Matter with India?" he argues that india's dysfunctional judicial system makes businesses to rely on the oldest form of trust, family, thus limiting innovation and scale. also very ironic, company building trust infrastructure for the internet (TLS) itself relying on the oldest form of trust, family. https://t.co/G4LcL6kABc

ca供应链 @S1r1u5_ source ↗

把漏洞利用原语搜索自动化,显著降低二进制利用的准入门槛。

yes, I just explained it. you filter based on the high IL expression that gives you the primitive you’re looking for. a vftable call is usually a this call, and typically you’ll have some control over that object. say you want the object to be freed and the reference count of the object is at +0x18. then the script would search for the expression *(this+0x18) -=1 , which decrements the ref counts to 0.

llm二进制利用方法 @chompie1337 source ↗

内核竞态是本地提权常见入口,讲解含可复用的利用思路。

RT @oct0xor: Alexander Popov (@a13xp0p0v) presents on exploiting a tricky race condition in the Linux kernel. This exploit is beautiful. #TheSAS2026 https://t.co/jx8kNXxTwZ

内核提权lpe议题 @0xocdsec source ↗

给出 AI 辅助原语挖掘的实现路径,思路可直接迁移到自有样本分析。

the script iterates each line of the function ripping out the high IL which looks like this. this could correspond to a primitive. sometimes the conditions behind the operation are not achievable, but restricting to functions that are small + the AI can usually reason ab reachability

llm二进制利用 @chompie1337 source ↗

RT @thsottiaux: Day 16/ You can now cache for 12h in the API and for logged in users across Codex &amp; Work we keep cache forever and you don’t have any penalty for coming back to an old conversation or resuming after a few hours and are now guaranteed to hit the cache.

llm工具 @weezerOSINT source ↗

不改 AD、不留日志的静默提权路径,BloodHound-CE 已能映射。

RT @g0h4n_0: 📝 New blog post - ESC13: Being a Phantom Member of a Group 👻🔥 Enroll in one cert template, and your certificate injects a group's SID into your Kerberos ticket. You "become" a member without ever being added to the group. No change in AD, no noise. 🦀 RustHound-CE v2.5.26 now maps this path in BloodHound-CE. Huge thanks to @marbasec for the feature and the many bug fixes! 🙏 📝 EN: https://t.co/7TgblR6pah 📝 FR: https://t.co/3lz7fs7oQF #ADCS #BloodHound #RedTeam #Rust #Pentest

adcs提权工具 @0xocdsec source ↗

显示 agent 编排已能产出真实可利用漏洞,本地 agent 风险面同步上升。

RT @J4ck3LSyN: @theonejvo It was a mix, I think between us 3 is was grok, glm, nemotron &amp; deepseek Ran in hermes-agent @acd421 @Archknight23 did i miss any?

ai_agentllm漏洞挖掘 @0xocdsec source ↗

自建证书透明监控成本低,可及时发现仿冒域名与新签发证书。

certstream-server-rust — High-performance Certificate Transparency (CT) monitoring tool written in Rust. Real-time stream of newly issued SSL/TLS certificates from CT logs... https://t.co/Nj0nSO0oir https://t.co/sBBc37gcsz

工具证书透明rust @KitPloit source ↗

RT @DevuanOrg: At Devuan, we're not driven by profit or marketing. We recognize the achievements of other communities through evidence, not hype. That's why we salute @OpenBSD on its 8.0 release: an outstanding community effort putting users' security first.

开源发布 @0xocdsec source ↗