Thanks @TenableSecurity for the mention! They used my php2shell PoC to reproduce CVE-2026-109618/109619 on several PHP 8.4/8.5 setups and wrote up why the exploit is so build-dependent: heap offsets, worker reuse, and why version checks lie. Great read.
https://t.co/8TY4OAPV6x
PHP2Shell update 🚀
13 → 56 perfiles.
5 → 30 con RCE.
Gracias a las aportaciones de la comunidad por ampliar la colección. Nuevos: stage-3 vessel 8.5.0–8.5.10, Debian 13/apxs2 y nginx+PHP-FPM.
https://t.co/jU6tLAhIDU
I'm glad to hear from threat intelligence that php2shell is not getting mass exploited. But you definitely should not believe you are safe. I have never said "panic turn it off". I am very much trying to help provide accurate (technical) information about the nature of the vulnerability. Which I hope threat intelligence takes into account.
But I have a problem with the communication, when some people claim the complexity makes it not realistic to exploit. That is just false. It must be clearly communicated, that vulnerable systems can be pwned (with a known vulnerability, not some theoretical 0day).
So as long as threat intelligence says this:
- we see no mass exploitation, you can relax a bit
- this is exploitable on real systems, you must patch when you can
we have common ground.
🚨 SON DAKİKA | PHP'de yeni RCE açığı: CVE-2026-115560
PHP, PHP2Shell yamalarından iki gün sonra yeni bir uzaktan kod çalıştırma (RCE) açığı duyurdu.
Bu açık, belirli bir session.serialize_handler yapılandırması ile session.upload_progress özelliğinin birlikte kullanılmasını gerektiriyor.
Önerilen geçici çözüm:
session.upload_progress.enabled = Off
Ayar, PHP2Shell (CVE-2026-109619) için de önerilen mitigasyonlardan biri. Daha önce uygulayıp kapalı tutanların ek işlem yapmasına gerek yok.
📌 CVE-2026-115560 için henüz etkilenen sürümler, CVSS skoru ve düzeltme içeren sürümler açıklanmadı. Bu açık için doğrulanmış saldırı bildirilmiş değil.
Not: 21 Ekim'de yayınlanan PHP2Shell yamaları bu yeni açığı düzeltmiyor.
İlgili duyuru: https://t.co/d5CdH5qdJY
https://t.co/ZDBbde7XfM
php2shell OOM in mod_php (CVE-2026-109618): the fake-HashTable fire only runs ~14-byte commands, too short for a shell. Poison a PHP session via upload_progress with <?php system(base64_decode())?>, then exec in 11 bytes: php /tmp/s* www-data shell. Full PoCs in my github repo!
#php2shell #CVE #InfoSec #PHP #RCE
Completely agree here - You should patch PHP but the sky is not falling yet. The issue here is going to be all the IOT Crap that just does not ever get updated. There still is time to patch.
If you want to see the state of things, find a cyber range (I prefer HackTheBox). Go look for a bunch of old PHP Boxes and try the exploit out on them. Right now I think you are going to have a tough time finding a retired box that you can exploit with PHP2SHELL.... Atleast I hope you do because I am having a hell of a time doing it and both love/hate to learn how much of a noob I am at times.
AI is really good at making crap look good. I know the POC that shows a POC of <?php echo("OK") ?> being exploited sounds scary. But when you look deeper there are a lot of hard-coded offsets that make it only work on that target.
#php2shell. seems to be hard to exploit (mem corruption vs aslr). yes, public pocs now are lab-only rces. this is a good one: https://t.co/ta8R61Aydj
limited to an unrealistic one-worker php-fpm. but it can be upgraded to N workers if we manage to hold N-1 workers :) here it is.
Awesome-Offensive-AI-Agentic-Landscape: a curated map of offensive AI research and tooling.
46 open-source projects, 11 models, 7 MCP servers, 116 papers, and 13 benchmarks covering autonomous pentesting, CTF agents, and red teaming.
Focuses on using AI for offensive security—not attacking AI systems. Includes documented curation and removal of out-of-scope projects.
https://t.co/sbwcbQTDkU
Claude-BugHunter provides a curated bundle of 83 skills for Claude Code designed to assist with vulnerability auditing, CVE tracking, and structured red team reporting.
https://t.co/4u8XsRFDIg
Run local uncensored AI Model for cybersecurity : that's actually built for red teamers & pentesting
- Trained to handle offensive & defensive cyber like a pro.
- Exploit crafting, vuln analysis, remediation, the works.
- Long context for digging through massive logs.
- And small enough to run consumers hardware.
Runs on a single GPU, thinks like a seasoned pentester, and won't lecture you about ethics.
This thing doesn't refuse it delivers. No more jailbreaking generic models just to get useful output.
If you're in InfoSec, DevSecOps, or just love playing with powerful open tools this deserves a spot in your arsenal.
THIS IS F**KING INSANE
i found an open source AI pentesting tool that deploys multiple agents to automate security testing.
and it does more than just scan.
PENTEST-SWARM-AI can:
→ discover subdomains and live hosts
→ scan ports and find vulnerabilities
→ validate findings with evidence
→ generate exportable reports
→ connect to Claude Desktop and Cursor
→ automate scans through GitHub
Actions
multiple AI agents. one security testing workflow.
only test systems you own or have permission to assess.
REPO BELOW
REA + Opus 5.5 is f*ucking insane…
I reverse engineered GTA 5 and Sonic with it, and an hour later Sonic was running through f*ucking Los Santos
One free repo lets your AI agent open any game on your PC, find the exact code behind every jump, crash and speed boost, and rebuild it your way.
Here's how to take apart your favorite game step by step:
step 1 → install REA in one prompt: npx rea-agents@latest setup, approve the plan, restart Claude Code
step 2 → pick your game and the route: .exe games go through Ghidra, Unity games read as .NET with no extra engine, Android games through JADX
step 3 → find the mechanic: tell REA to search the game for words like speed, jump or score. It lands on the exact function that controls them
step 4 → trace it: REA shows who calls that function and what it changes, so you see the whole chain from button press to what happens on screen
step 5 → prove it live: play for 60 seconds while REA watches, and it confirms the rule actually fires
step 6 → write it down: every rule goes into SPEC.md with the number, the function and the proof. No guessing
step 7 → rebuild it your way: Claude turns SPEC.md into real Godot code with your own assets, then tests it against the original until it matches
That's how I crossed Sonic with GTA 5. Same speed, same physics feel, running on streets he was never supposed to see.
Your favorite game stops being a black box and becomes a playground.
Save it and try it with your favorite game. Before u start read the article below↓