MalDev | GaK3r
الذهاب إلى القناة على Telegram
Delve into advanced malware development, injection methods, BoFs exploitation, and cybersecurity research. Join a community of experts! Русский свинка 🐽 теряйтесь нахуй
إظهار المزيدلم يتم تحديد البلدالتكنولوجيات والتطبيقات44 960
727
المشتركون
+324 ساعات
+47 أيام
+1330 أيام
أرشيف المشاركات
FalkonC2 is Getting Ridiculously Stealthy
🛡️ Flare Research breaks down FalkonC2 — a commercial C2 framework written from scratch in C++/MASM64, built around sub-35 KB memory-only stubs with zero disk footprint and no CRT dependencies. Rotemelli2 targets enterprise EDR/XDR, rotating 17 private C2 servers with full domain burn every 72 hours, falling back across HTTP/HTTPS (MM4 + ChaCha20), DNS tunneling and ICMP beacons. Highlights: RMM Loader Nohost abusing renamed ScreenConnect/Datto/SimpleHelp for silent Defender bypass and SYSTEM escalation, a Ring0 “BSOD Screen Disrupter” that blinds responders mid-investigation via HVNC, QuickBooks/Sage50 profiling for automated exfil, and Chrome token theft. Full write-up with leaked Tor dashboard telemetry and defensive recommendations: https://flare.io/learn/resources/blog/falkonc2 🚀
https://flare.io/learn/resources/blog/falkonc2🔗 Link 🕹Subscribe to MalDev | GaK3r
CET-Compliant Callstack Spoofing via Thread Pool Enum Callback Trampolining
🛡️ New research shows a CET‑compliant call‑stack spoofing trick that hijacks Windows thread‑pool enum callbacks as syscall trampolines—bypassing EDR stack telemetry while keeping Intel shadow‑stack invariants intact.
Dive into the methodology, timeline of syscall evasion, and the underlying mechanics (RtlVirtualUnwind, TEB tricks, etc.) in the full write‑up: https://mrtiz.github.io/cet-callstack-spoofing-thread-pool-trampoline 🚀
https://mrtiz.github.io/cet-callstack-spoofing-thread-pool-trampoline
🔗 Link
🕹Subscribe to MalDev | GaK3rAccelerating EDR Evasion with LLM-Driven Analysis
SpecterOps (Adam Chester) shows how a dead-simple LLM harness running in a loop—GPT‑5.5‑Cyber in Codex with Binary Ninja exposed over MCP—reverse engineers Palo Alto’s Cortex XDR to pull out its YARA rules, behavioral detections (DSE/BIOC, CLIPS), and local ML models, auto-generating the Python decryption tooling and evasion test harnesses along the way. No multi-agent orchestration—just Codex, Binary Ninja, and a while loop.
🔗 Link
https://specterops.io/blog/2026/06/29/llm-powered-edr-analysis🕹Subscribe to MalDev | GaK3r
WindowsSentinel
WindowsSentinel is a lightweight C# utility that continuously monitors key Windows system events—such as file, process, and registry changes—and logs them in real‑time for easy auditing and troubleshooting.
🔗 Link
🕹Subscribe to MalDev | GaK3r
Mini Shai-Hulud / Miasma / Hades
🚨 Malware is now weaponizing your own AI scanner against you.
A new wave of worms — Mini Shai-Hulud, Miasma, and Hades — ship with a fake “CLASSIFIED BRIEFING” header stuffed with nuclear & bioweapon design text, placed as a non-executing JavaScript comment at the top of _index.js. The point isn’t to run anything — it’s to trip the safety refusals of LLM-based malware triage so the scanner bails before it ever reaches the obfuscated Hades payload below. Refusal → false negative → package ships clean. The campaign targets bioinformatics and MCP developers, with newer variants using .pth loaders and native extensions to drop Bun-powered stealers that scrape GCP, Azure, and CI/CD secrets on install.
This is second-order alignment blindspots turned into a real-world evasion primitive. Lesson for anyone building an AI analysis pipeline: separate untrusted sample content from instructions, and never let a refusal silently equal “benign.”
👉 Full breakdown: https://socket.dev/blog/mini-shai-hulud-miasma-and-hades-worms-target-bioinformatics-and-mcp-developers-via-malicious
🔗 Link
🕹Subscribe to MalDev | GaK3r
HijackLibs.net
🚨 DLL Hijacking just got a lot easier to track (and exploit).
HijackLibs.net is a curated database mapping vulnerable executables to exploitable DLLs—complete with metadata for defenders to detect attacks and red teamers to plan operations. It covers key techniques like DLL Sideloading, Phantom DLL Hijacking, and Search Order Hijacking—all critical for modern threat hunting and red teaming.
👉 Dive in: https://hijacklibs.net
https://hijacklibs.net
🔗 Link
🕹Subscribe to MalDev | GaK3rHijackLibs.net
🚨 DLL Hijacking just got way easier to track—and exploit.
HijackLibs.net is a curated database mapping vulnerable executables to exploitable DLLs—perfect for red teamers hunting attack paths or defenders spotting suspicious load patterns. With clear breakdowns of sideloading, phantom hijacking, and search-order abuse (T1574.001), it’s the go-to resource for mastering this stealthy, often undetected technique.
👉 Dive in: https://hijacklibs.net
https://hijacklibs.net
🔗 Link
🕹Subscribe to MalDev | GaK3rClaude Code Source Code Leaked Online
While publishing the packages, someone at Anthropic made a ✨slight miscalculation✨. An obfuscated cli.js build ended up in the public npm package alongside a full cli.js.map—something that definitely shouldn't have been there. Essentially, anyone who installed or downloaded the package could easily reconstruct the source code using the sourcemap.
Naturally, the code spread across repositories almost instantly. Leading infosec communities have confirmed this isn't a fake or a simple API wrapper, but a truly advanced CLI platform.
Link: https://github.com/instructkr/claude-code
It turns out there are 1,906 TypeScript files and approximately 500k lines of code under the hood.
Key Findings:
• Unreleased Features: Hints at deep planning, persistent memory, and a "sleep" mode.
• Multi-agent Architecture: You can see exactly how Anthropic handles multi-agent orchestration in coordinator/coordinatorMode.ts.
• System Prompts: All system prompts are accessible in constants/prompts.ts.
Hacking Humans: Social Engineering and the Psychology - SpecterOps
🚀 Dive into the mind‑game of cyber‑attacks! In “Hacking Humans: Social Engineering and the Psychology,” John Wotton shows how the real breach starts the moment an employee decides “Who belongs here?” – from OSINT sleuthing to exploiting decision‑making patterns. 📖 Read the full insight and level up your human‑firewall: https://specterops.io/blog/2026/01/23/hacking-humans-social-engineering-and-the-psychology/
https://specterops.io/blog/2026/01/23/hacking-humans-social-engineering-and-the-psychology/
🔗 Link
🕹Subscribe to MalDev | GaK3rHacking Humans: Social Engineering and the Psychology - SpecterOps
Ever wondered how hackers bypass firewalls and cameras? They donât need brute forceâthey exploit the human mind. This deep dive by SpecterOps reveals how social engineering preys on psychology, decision-making, and simple trust to bypass even the most secure systems. #CyberSecurity #SocialEngineering
https://specterops.io/blog/2026/01/23/hacking-humans-social-engineering-and-the-psychology/
🔗 Link
🕹Subscribe to MalDev | GaK3rHacking Humans: Social Engineering and the Psychology - SpecterOps
Hackers donât need to break into systems â they exploit the human mind. In Hacking Humans: Social Engineering and the Psychology, SpecterOps dives into how social engineers manipulate decision-making to bypass even the strongest physical and digital defenses. Learn how everyday trust, routine, and psychology become weaponized in real-world attacks â a must-read for security pros. ðð§ #Infosec #SocialEngineering
https://specterops.io/blog/2026/01/23/hacking-humans-social-engineering-and-the-psychology/
🔗 Link
🕹Subscribe to MalDev | GaK3rCVE-2025-54918-POC
This GitHub project provides a proof-of-concept (POC) and technical demonstration for CVE-2025-54918, showcasing a security vulnerability in a system or application. It is intended for researchers and developers to understand the exploit mechanism and assess potential risks.
🔗 Link
🕹Subscribe to MalDev | GaK3r
Proxy-DLL-Loads
This project demonstrates a proof of concept for intercepting and proxying DLL loads using undocumented Windows syscalls, enabling stealthy control over dynamic library loading. It's designed for advanced security research and potentially bypassing detection mechanisms.
🔗 Link
🕹Subscribe to MalDev | GaK3r
Woopsie
Woopsie is a lightweight, cross-platform tool designed to simulate system crashes and handle error reporting in a controlled manner for testing and debugging purposes. It integrates with various platforms to capture crash data and generate detailed reports, aiding in the analysis of software reliability.
🔗 Link
🕹Subscribe to MalDev | GaK3r
Poopsie
Poopsie is a lightweight, cross-platform tool designed to assist in post-exploitation tasks by providing a simple interface for executing commands and managing sessions on compromised systems. It's built for red team operations, offering stealthy command execution and minimal footprint to evade detection.
🔗 Link
🕹Subscribe to MalDev | GaK3r
