Remote Command Execution in Google Cloud with Single Directory Deletion Introduction Hello, I’m RyotaK (@ryotkak ), a security engineer at GMO Flatt Security Inc. A while ago, I participated in the Google Cloud VRP bugSWAT, a live hacking event organized by Google. During...

𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗪𝗼𝗿𝘁𝗵 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 - 𝗪𝗲𝗲𝗸 𝟭𝟯, 𝟮𝟬𝟮𝟲
Only one entry but definitely worth reading!

☁️ 𝗥𝗲𝗺𝗼𝘁𝗲 𝗖𝗼𝗺𝗺𝗮𝗻𝗱 𝗘𝘅𝗲𝗰𝘂𝘁𝗶𝗼𝗻 𝗶𝗻 𝗚𝗼𝗼𝗴𝗹𝗲 𝗖𝗹𝗼𝘂𝗱 𝘄𝗶𝘁𝗵 𝗦𝗶𝗻𝗴𝗹𝗲 𝗗𝗶𝗿𝗲𝗰𝘁𝗼𝗿𝘆 𝗗𝗲𝗹𝗲𝘁𝗶𝗼𝗻
This one is a real tour de force: flatt.tech/research/pos....

🧑‍🎓 Learning of the day for me, once again thanks to @pentesterlab.com (for the presentation of the behavior and the code review lab) and Claude (for the detailed explanation).

#appsec #appsecurity

💀 𝗖𝗧𝗙 𝗶𝘀 𝗱𝗲𝗮𝗱*
A good rant on the impact of AI on CTF... k3ng.xyz/blog/ctf-is-....

🐧 𝗦𝗮𝘀𝗵𝗶𝗸𝗼
Sashiko is an agentic Linux kernel code review system that monitors public mailing lists to thoroughly evaluate proposed Linux kernel changes. sashiko.dev.

Hyoketsu – Solving the Vendor Dependency Problem in RE › Searchlight Cyber The Long Decompilation Process Over the last eight or so years of performing security research at Assetnote, our research team has looked at countless enterprise applications that ship with hundreds, ...

🛠️ 𝗛𝘆𝗼𝗸𝗲𝘁𝘀𝘂 – 𝗦𝗼𝗹𝘃𝗶𝗻𝗴 𝘁𝗵𝗲 𝗩𝗲𝗻𝗱𝗼𝗿 𝗗𝗲𝗽𝗲𝗻𝗱𝗲𝗻𝗰𝘆 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 𝗶𝗻 𝗥𝗘
Reversing Java and C# applications just became a lot easier thanks to the SearchLight Cyber team (ex: Assetnote): slcyber.io/research-cen....

𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗪𝗼𝗿𝘁𝗵 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 - 𝗪𝗲𝗲𝗸 𝟭𝟮, 𝟮𝟬𝟮𝟲
AI doing research, AI killing CTF

🤖 𝗧𝗲𝘀𝘁𝗶𝗻𝗴 𝗔𝗜 𝗳𝗼𝗿 𝗩𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵: 𝟰 𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵𝗲𝘀 & 𝗪𝗵𝗲𝗿𝗲 𝗜 𝗙𝗮𝗶𝗹𝗲𝗱
If you can only read one thing this week, make it this article: xclow3n.github.io/post/7.

Example of execution.

🧑‍🎓 Learning of the day for me thanks to @pentesterlab.com (for the presentation of the behavior and the code review lab) and Claude (for the detailed explanation):

#appsec #appsecurity

How "Strengthening Crypto" Broke Authentication: FreshRSS and bcrypt's 72-Byte Limit As part of our CVE monitoring, we came across GHSA-pcq9-mq6m-mvmp (CVE-2025-68402), an authentication bypass in FreshRSS, a self-hosted RSS aggregator. It ...

A commit meant to "strengthen the crypto" in FreshRSS ended up removing the need for a correct password.

Why?
Longer SHA-256 nonce + bcrypt truncation at 72 bytes.

A nice example of why secure systems are about composition, not just stronger primitives.

pentesterlab.com/blog/freshrs...

CVE-2026-29000: Critical Auth Bypass in pac4j-jwt: Full PoC Using Only a Public Key CodeAnt AI found a critical authentication bypass in pac4j-jwt where an attacker can impersonate any user using only the RSA public key. Full PoC and disclosure.

⛓️‍💥 𝗔𝘂𝘁𝗵𝗲𝗻𝘁𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗕𝘆𝗽𝗮𝘀𝘀 𝗶𝗻 𝗽𝗮𝗰𝟰𝗷
Another issue with a library leveragining JWT: www.codeant.ai/security-res....

😎 𝗩𝗶𝗯𝗲 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗥𝗮𝗱𝗮𝗿
A cool little project to track the security issues created by vibe coding: vibe-radar-ten.vercel.app.

✨ 𝗧𝗵𝗲 𝗠𝗖𝗣 𝗔𝘂𝘁𝗵𝗡/𝗭 𝗡𝗶𝗴𝗵𝘁𝗺𝗮𝗿𝗲
A reminder of the mess AuthN/Z with MCP is: blog.doyensec.com/2026/03/05/m....

mitmproxy for fun and profit: Interception and Analysis of Application mitmproxy for fun and profit: Interception and Analysis of Application

🚥 𝗺𝗶𝘁𝗺𝗽𝗿𝗼𝘅𝘆 𝗳𝗼𝗿 𝗳𝘂𝗻 𝗮𝗻𝗱 𝗽𝗿𝗼𝗳𝗶𝘁: 𝗜𝗻𝘁𝗲𝗿𝗰𝗲𝗽𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗔𝗻𝗮𝗹𝘆𝘀𝗶𝘀 𝗼𝗳 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗧𝗿𝗮𝗳𝗳𝗶𝗰
A write-up on how to use mitmproxy: www.synacktiv.com/en/publicati....

𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗪𝗼𝗿𝘁𝗵 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 - 𝗪𝗲𝗲𝗸 𝟭𝟬, 𝟮𝟬𝟮𝟲
A great mix of content this week!

🔒 𝗜𝗿𝗼𝗻𝗖𝘂𝗿𝘁𝗮𝗶𝗻: 𝗔 𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝗹 𝗔𝗜 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁 𝗕𝘂𝗶𝗹𝘁 𝗦𝗲𝗰𝘂𝗿𝗲 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝗚𝗿𝗼𝘂𝗻𝗱 𝗨𝗽
Niels Provos (from OpenBSD's systrace) is sharing a new tool to sandbox your AI assistant: www.provos.org/p/ironcurtai....

vinext: Vibe-Hacking Cloudflare's Vibe-Coded Next.js Replacement Cloudflare built a Next.js replacement in a week with AI for $1100. We pointed Hacktron at it to find what the tests missed.

⛈️ 𝘃𝗶𝗻𝗲𝘅𝘁: 𝗩𝗶𝗯𝗲-𝗛𝗮𝗰𝗸𝗶𝗻𝗴 𝗖𝗹𝗼𝘂𝗱𝗳𝗹𝗮𝗿𝗲'𝘀 𝗩𝗶𝗯𝗲-𝗖𝗼𝗱𝗲𝗱 𝗡𝗲𝘅𝘁.𝗷𝘀 𝗥𝗲𝗽𝗹𝗮𝗰𝗲𝗺𝗲𝗻𝘁
It's raining bugs in the cloud. A great example of agent capabilities on a never-seen-before target: www.hacktron.ai/blog/hacking....

100+ Kernel Bugs in 30 Days High-Scale Driver Vulnerability Research with Agent Swarms

🪟 𝟭𝟬𝟬+ 𝗞𝗲𝗿𝗻𝗲𝗹 𝗕𝘂𝗴𝘀 𝗶𝗻 𝟯𝟬 𝗗𝗮𝘆𝘀
Behind the (impressive) result, the methodology is probably the most important. Make sure you read between the lines: substack.com/home/post/p-....

𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗪𝗼𝗿𝘁𝗵 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 - 𝗪𝗲𝗲𝗸 𝟵, 𝟮𝟬𝟮𝟲
Mostly AI...

💻 𝗕𝗿𝗼𝘄𝘀𝗲𝗿-𝗕𝗮𝘀𝗲𝗱 𝗣𝗼𝗿𝘁 𝗦𝗰𝗮𝗻𝗻𝗶𝗻𝗴 𝗶𝗻 𝘁𝗵𝗲 𝗔𝗴𝗲 𝗼𝗳 𝗟𝗡𝗔
Leveraging Local Network Access to create a port scanner! wiki.notveg.ninja/tools/lna-po....

PentesterLab: Learn with our JavaScript Code Review The JavaScript Code Review Badge is our badge dedicated to security code review in JavaScript. It covers the discovery of weaknesses and vulnerabilities using source code review.

6 new code review labs just dropped 🚀
+3 for JavaScript Code Review
+3 for Python Code Review

JS: pentesterlab.com/badges/javas...

Python: pentesterlab.com/badges/pytho...

CVE Archeologist's field guide Methodology and Lessons from 10 Vulnerability Analyses Written by Louis Nyffenegger

As I mentioned last week I ordered some books (which got delivered this week)
CVE ARCHEOLOGIST'S FIELD GUIDE by Louis Nyffenegger @pentesterlab.com

Using threat modeling and prompt injection to audit Comet Trail of Bits used ML-centered threat modeling and adversarial testing to identify four prompt injection techniques that could exploit Perplexity’s Comet browser AI assistant to exfiltrate private Gma...

🤖 𝗨𝘀𝗶𝗻𝗴 𝘁𝗵𝗿𝗲𝗮𝘁 𝗺𝗼𝗱𝗲𝗹𝗶𝗻𝗴 𝗮𝗻𝗱 𝗽𝗿𝗼𝗺𝗽𝘁 𝗶𝗻𝗷𝗲𝗰𝘁𝗶𝗼𝗻 𝘁𝗼 𝗮𝘂𝗱𝗶𝘁 𝗖𝗼𝗺𝗲𝘁
The team at Trail of Bits is sharing some key learnings from their audit of Comet (AI browser) blog.trailofbits.com/2026/02/20/u....

HeroDevs Blog | CVE-2026-0603: Second-Order SQL Injection in Hibernate UPDATE/DELETE (InlineIdsOrClauseBuilder) CVE-2026-0603 is a high-severity second-order SQL injection vulnerability in Hibernate ORM caused by unsafe ID inlining during UPDATE and DELETE operations. Learn impact, root cause, and remediation.

☕️ 𝗖𝗩𝗘-𝟮𝟬𝟮𝟲-𝟬𝟲𝟬𝟯: 𝗦𝗲𝗰𝗼𝗻𝗱-𝗢𝗿𝗱𝗲𝗿 𝗦𝗤𝗟 𝗜𝗻𝗷𝗲𝗰𝘁𝗶𝗼𝗻 𝗶𝗻 𝗛𝗶𝗯𝗲𝗿𝗻𝗮𝘁𝗲 𝗨𝗣𝗗𝗔𝗧𝗘/𝗗𝗘𝗟𝗘𝗧𝗘 (𝗜𝗻𝗹𝗶𝗻𝗲𝗜𝗱𝘀𝗢𝗿𝗖𝗹𝗮𝘂𝘀𝗲𝗕𝘂𝗶𝗹𝗱𝗲𝗿)
A bit of a stretch but an interesting insight into Hibernate: www.herodevs.com/blog-posts/c....

😱 𝗩𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗗𝗶𝘀𝗰𝗹𝗼𝘀𝘂𝗿𝗲: 𝗝𝗪𝗧 𝗔𝘂𝘁𝗵𝗲𝗻𝘁𝗶𝗰𝗮𝘁𝗶𝗼𝗻 𝗕𝘆𝗽𝗮𝘀𝘀 𝗶𝗻 𝗢𝗽𝗲𝗻𝗜𝗗 𝗖𝗼𝗻𝗻𝗲𝗰𝘁 𝗔𝘂𝘁𝗵𝗲𝗻𝘁𝗶𝗰𝗮𝘁𝗼𝗿 𝗳𝗼𝗿 𝗧𝗼𝗺𝗰𝗮𝘁
The exact same vulnerability I found in HarbourJWT but in a much cooler target, still not fixed... insinuator.net/2026/02/jwt-....

Almost Impossible: Java Deserialization Through Broken Crypto in OpenText Directory Services › Searchlight Cyber Introduction We recently found ourselves looking into OpenText Directory Services (OTDS). We had seen it present on our customer's attack surface, and it seemed to be an interesting target. OTDS is a ...

☕️ 𝗔𝗹𝗺𝗼𝘀𝘁 𝗜𝗺𝗽𝗼𝘀𝘀𝗶𝗯𝗹𝗲: 𝗝𝗮𝘃𝗮 𝗗𝗲𝘀𝗲𝗿𝗶𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗧𝗵𝗿𝗼𝘂𝗴𝗵 𝗕𝗿𝗼𝗸𝗲𝗻 𝗖𝗿𝘆𝗽𝘁𝗼 𝗶𝗻 𝗢𝗽𝗲𝗻𝗧𝗲𝘅𝘁 𝗗𝗶𝗿𝗲𝗰𝘁𝗼𝗿𝘆 𝗦𝗲𝗿𝘃𝗶𝗰𝗲𝘀
What an adventure in Java Deserialisation... slcyber.io/research-cen....

𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗪𝗼𝗿𝘁𝗵 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 - 𝗪𝗲𝗲𝗸 𝟴, 𝟮𝟬𝟮𝟲
Java x2, Go, JWT and a sprinkling of AI

🦫 𝗖𝗧𝗙𝘁𝗶𝗺𝗲.𝗼𝗿𝗴 / 𝗷𝘂𝘀𝘁𝗖𝗧𝗙 [*] 𝟮𝟬𝟮𝟬 / 𝗚𝗼-𝗳𝘀 / 𝗪𝗿𝗶𝘁𝗲𝘂𝗽
A cool Golang quirk via an unintended CTF solution ctftime.org/writeup/25852.

New lab: CVE-2026-24895 — FrankenPHP Path Confusion RCE (Unicode)

People think "lowercase it" is harmless. In Unicode it’s not.

Case folding can do weird mappings (Turkish i, Kelvin sign…), and sometimes worse: UTF-8 byte length changes.

Hands-on lab: pentesterlab.com/exercises/cv...

🚛 𝗧𝗿𝗮𝗶𝗹𝗶𝗻𝗴 𝗗𝗮𝗻𝗴𝗲𝗿: 𝗲𝘅𝗽𝗹𝗼𝗿𝗶𝗻𝗴 𝗛𝗧𝗧𝗣 𝗧𝗿𝗮𝗶𝗹𝗲𝗿 𝗽𝗮𝗿𝘀𝗶𝗻𝗴 𝗱𝗶𝘀𝗰𝗿𝗲𝗽𝗮𝗻𝗰𝗶𝗲𝘀
Probably one of the lesser-known features of HTTP... Trailers. This post provides details on the feature and explains how they can be leveraged to find security issues. www.sebsrt.xyz/blog/trailin....

RCE in Google's AI code editor Antigravity - $10000 Bounty Hacktron AI Research Team discovered a critical RCE in Google’s Antigravity IDE that lets attackers take over your system just by opening a malicious website.

🤖 𝗥𝗖𝗘 𝗶𝗻 𝗚𝗼𝗼𝗴𝗹𝗲'𝘀 𝗔𝗜 𝗰𝗼𝗱𝗲 𝗲𝗱𝗶𝘁𝗼𝗿 𝗔𝗻𝘁𝗶𝗴𝗿𝗮𝘃𝗶𝘁𝘆 - $𝟭𝟬𝟬𝟬𝟬 𝗕𝗼𝘂𝗻𝘁𝘆
A detailed blog post on hacking Antigravity with a lot of interesting details on its inner workings. www.hacktron.ai/blog/hacking....

🤺 𝗪𝗵𝗲𝗻 𝗧𝘄𝗼 𝗣𝗮𝗿𝘀𝗲𝗿𝘀 𝗗𝗶𝘀𝗮𝗴𝗿𝗲𝗲: 𝗘𝘅𝗽𝗹𝗼𝗶𝘁𝗶𝗻𝗴 𝗤𝘂𝗲𝗿𝘆 𝗦𝘁𝗿𝗶𝗻𝗴 𝗗𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁𝗶𝗮𝗹𝘀 𝗳𝗼𝗿 𝗫𝗦𝗦
If you enjoy parser differential issues, you are going to love this exploit. Don't think "It's a CTF challenge" or "It's just an XSS", read between the lines to find the real gold. blog.voorivex.team/when-two-par....

Introducing Augustus: Open Source LLM Prompt Injection Tool Explore LLM fingerprinting and its role in exposing vulnerable Ollama servers online to unauthorized access risks.

⚒️ 𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗔𝘂𝗴𝘂𝘀𝘁𝘂𝘀: 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲 𝗟𝗟𝗠 𝗣𝗿𝗼𝗺𝗽𝘁 𝗜𝗻𝗷𝗲𝗰𝘁𝗶𝗼𝗻 𝗧𝗼𝗼𝗹
Praetorian is back with another tool (one of their 12 Caesars): Augustus... Make sure you check it out! www.praetorian.com/blog/introdu....

Breaking Down CVE-2026-25049: How TypeScript Types Failed n8n's Security | hetmehta.com Deep technical analysis of CVE-2026-25049: How type confusion bypassed n8n's security patch and why TypeScript types aren't runtime security boundaries

⨐ 𝗕𝗿𝗲𝗮𝗸𝗶𝗻𝗴 𝗗𝗼𝘄𝗻 𝗖𝗩𝗘-𝟮𝟬𝟮𝟲-𝟮𝟱𝟬𝟰𝟵: 𝗛𝗼𝘄 𝗧𝘆𝗽𝗲𝗦𝗰𝗿𝗶𝗽𝘁 𝗧𝘆𝗽𝗲𝘀 𝗙𝗮𝗶𝗹𝗲𝗱 𝗻𝟴𝗻'𝘀 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆
A great explanation of the recent vulnerabilities impacting n8n. If you are working in security on TypeScript projects, it's a must read. hetmehta.com/posts/n8n-ty....

𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗪𝗼𝗿𝘁𝗵 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 - 𝗪𝗲𝗲𝗸 𝟳, 𝟮𝟬𝟮𝟲
Parser Differential, TypeScript and AI

👇