Quang Nguyen retweeted
Learn how to use Frida to bypass an iOS app's encrypted database, not by breaking the encryption, but by intercepting the passphrase at runtime and querying the unlocked database directly from a script: 8ksec.io/advanced-frida-usag…
Follow @8kSec for more hands-on security content
Quang Nguyen retweeted
CVE-2026-35273: SSRF to RCE in Oracle PeopleSoft. ZDI's @bobbygould5 provides a deep dive into the bug(s) used in the recent PeopleSoft attacks. His analysis includes root cause and video demo. zerodayinitiative.com/blog/2…
Quang Nguyen retweeted
OAuth - More OBroken Than You Think
OBroken is here. The first tool to autonomously find, exploit and prove OAuth authorization bypasses end to end.
Our proprietary OAuth Authorization Bypass Scanner, covers 23 techniques across 17 vulnerability classes, each confirmed with a standalone Find/Forge/Fire PoC ready to report and collect.
Part of the Broken Token series alongside jBroken, our JWT scanner with 40+ techniques. Same methodology, same output format, different protocol layer.
The screenshot below shows oBroken running against the Missing aud Validation test case (CVE-2025-9803) on our OAuth testbed.
The tool scrapes the client app, discovers the AS, requests a token bound to an attacker-controlled resource, presents it to the real API and confirms it's accepted without audience validation.
One URL in, full PoC out.
Covers response mode escape, state manipulation, PKCE downgrade variants, DCR abuse, race conditions and aud validation, among others.
Running on our lab terminal today, landing on an private API someday. If you have interest, let us know.
Testbed below.
OAuth Bypass Testbed
403.brutelogic.net/authz/oau…
Test your skills against a LOGIN (randomly vulnerable) and/or 17 AUTHENTICATION SERVER endpoints.
Like with JWT (JBroken), our fully automated OAuth tool (OBroken) with complete FFF (Find, Forge, Fire) PoCs, aces this testbed.
Quang Nguyen retweeted
‼️ A critical Signal bug could keep your microphone, and possibly your camera, live after you hang up.
The other caller can flood the call's event queue so your hang-up is never processed.
Dubed CVE-2026-113380 it scored 9.3 and got fixed in September.
Quang Nguyen retweeted
Do you know who illustrated the comic on calif.io/php2shell?
Her name is Stephanie Cortes. And her husband is the hacker who discovered php2shell: @lcfr_hax.
This artist-hacker couple has the best story, stay tuned!
Quang Nguyen retweeted
Check out our latest deep-dive for #AhsayCBS, an enterprise backup solution.
Reported as exploited in-the-wild by @HuntressLabs, CVE-2026-105133 and CVE-2026-105134 allow an unauthenticated attacker to execute commands as SYSTEM or root on the server.
horizon3.ai/attack-research/…
Quang Nguyen retweeted
🚨 PoC published for a critical unauthenticated Splunk Enterprise remote code execution vulnerability; CVE-2026-76268
PoC: github.com/murrez/CVE-2026-7…
The flaw affects the Patroni REST API used by the PostgreSQL sidecar on Splunk Enterprise search head cluster members. Critical configuration operations can be accessed without authentication, allowing an attacker with network access to the Patroni service to modify PostgreSQL configuration and execute operating system commands.
CVSS: 9.8 Critical
Affected: Splunk Enterprise 10.2.0 - 10.2.6 and 10.4.0 - 10.4.2
Fixed: Splunk Enterprise 10.2.7 and 10.4.3
The PoC includes vulnerability checking, Patroni port discovery, multi-target scanning, out-of-band verification, and an exploit mode demonstrating the RCE chain.
Quang Nguyen retweeted
Auto-patch for OpenAnt: Generating trust in autonomous patching knostic.ai/blog/auto-patch-f…
Quang Nguyen retweeted
Oh damn! It's already live.
We’re expanding our Cyber Verification Program to give security professionals broader access to our most capable models.
Through this program, verified security professionals can access Claude Mythos 5.1, Opus 5.5, and Sonnet 5.5 with safeguards designed for defensive work.
We’re also opening up new tiers to allow for authorized offensive work, like penetration testing and red-teaming.
anthropic.com/news/cyber-ver…
Quang Nguyen retweeted
Perplexity wins on Hugging Face Decision Index (benchmark for decision models). Our model is open-weights and accessible to all!
pplx-decider-v1.1-27b, our updated open weights multimodal decision model, is now available.
It scores the highest on the new @huggingface Decision Index 0.3 benchmark. pplx-decider-v1.1-27b costs half as much as v1, at $0.02 per million input tokens.
huggingface.co/spaces/multim…
Quang Nguyen retweeted
Adminer 0-days 🤯🔥
1️⃣ Pre-Auth RCE (MSSQL DSN injection) → web shell, no login
2️⃣ Stored XSS (rogue MySQL + CSP nonce bypass) → session hijack
3️⃣ Auth RCE (SQLite VACUUM INTO) → root shell
Blocklists fail. Every time.
Fix: 5.4.3+
🔗 blog.voorivex.team/three-0-d…
👉 t.me/luckyhacker42
#BugBounty #InfoSec #RCE
Quang Nguyen retweeted
Crazy numbers!! 🔥🔥
Qwen 3.8 Flash Next on 2x M5 Ultra Mac Studios
~271 t/s decode
8000 tok/s PP
All you need is TensorFolder with MCDMA, connected with 2 bonded TB cables.
THIS IS NEXT LEVEL!
@Alibaba_Qwen Qwen 3.8 Next Flash 6 bit running at 400+tps peak with an average of ~271 tks decode and 8000tps PP on two @Apple M5U studios.
TensorFold running on both with MCDMA 2 bonded TB5 cables.
Concurrency is being measured.
@digitalix Thanks again for letting me access your two Studios. GLM5.3 Flash inbound 👀
LET'S GOOO 🚀
Quang Nguyen retweeted
Claude now works inside Google Docs, Sheets, and Slides, and those files also open inside Claude.
In Google Workspace, Claude sits in a sidebar next to your file, reads what you have open, and edits it in place. You can approve each edit before it lands.
Quang Nguyen retweeted
🚨 Uber 内部用来盯 AI agent 的那套安全系统 ADR，开源了，能把公司里没人报备就偷偷用上的 AI 工具揪出来。
GitHub 上 1623 star、172 个 fork，相关论文被 MLSys 2026 接收，系统本身已经在 Uber 生产环境跑着。
现在公司里的 agent 安全基本是个黑洞：谁在用 Cursor、谁装了 Claude Code、哪个客服 agent 被人用提示词绕过去了，安全团队看不见。ADR 把这件事拆成五块来做——先把没批准的 AI 工具找出来，再观测 agent 到底调了什么工具、执行了哪些动作，然后评测防御能力、检测威胁、拦住不安全的操作。
观测这块已经在生产环境里跨 macOS、Linux、Windows 采了七种以上 AI 编程工具的轨迹；检测用的是两层结构，先高召回粗筛，再对可疑会话做深一层的推理。它还带了一个 ADR-Bench，300 多个任务、134 个 MCP server，把 17 种 agent 攻击手法全覆盖了一遍。
要说清楚的是，这次开源的是 Discovery、Sensor、ADR-Bench 和 Detector 这几块，负责拦截的 Prevention 组件和离线的 Explorer 引擎没放出来。
agent 进了生产，安全这一课迟早得补。
GitHub：github.com/uber/ADR
Quang Nguyen retweeted
EndpointSecurity code appears in the release kernel. (iOS 27.2, 24B5099f)
com.apple.iokit.AppleSecurityResearch
esk_ios_research: device is not in research mode; kext inert
github.com/blacktop/ipsw-dif…
Quang Nguyen retweeted
Someone is porting PS5 games to run natively on PC.
No emulator. AnyPS5 converts the game's code into a normal Windows or Linux program.
A real game has already booted to the main menu and into gameplay with sound.
github.com/boykopovar/AnyPS5
Quang Nguyen retweeted
Cloudflare's Auto Router is out in public beta.
Set your model to `cloudflare/auto` and AI Gateway picks a model for each request based on the complexity of the task. It's also conscientious about reusing the prompt cache. Pretty slick.
Docs: developers.cloudflare.com/ai…
