mgeeky | Mariusz Banach

☢️ 2026 started with a bang! The project I've been building for the last three years - an Initial Access framework letting us weaponize 100+ file formats - is joining Outflank's OST & I'm joining too!😍 🔥 Same mission, now with joint R&D to ship even more high-quality RT tools!

Welcome @_RastaMouse to the gang! 😎 Exciting times ahead! 🔥🍻

I’ve been grinding hard on AI for the better part of the last 8+ months - learning, building, adapting, and pulling late nights just like so many others right now. Cutting through the FUD and hype, there is real potential here. Industry-breaking potential. The era we’ve been waiting for - to finally supercharge and develop the tools and platforms we’ve wanted to build for years - is here, and agent assistance is accelerating everything. With coding agents, I’ve built solid tools and had research breakthroughs that would have taken weeks or months before. These should feel like real wins worth celebrating. But honestly? I don’t feel victorious. In many ways, it just feels necessary to keep pace. As Dave said: adapt or be left behind - and for good reason. I’m not ready to be left behind. But damn, I’m tired. I’m tired of constantly reinventing myself. Tired of constantly re-tooling. Tired of the endless cycle of keeping up, the late nights, and the personal sacrifices that come with it. I’ve even lost the desire to share knowledge and research with the community the way I used to. From the conversations I’ve had, I’m far from alone - many others in this space feel the same but don’t necessarily vocalize it outside of smaller circles. Is it because I see AI purely as a threat? Not really. The offensive side of our industry has been heading this way for a while, and I’ve been moving with it. The truth is, the excitement Dave describes is real - but for me right now, it’s mixed with exhaustion. I’m grateful for the breakthroughs, yet I catch myself wondering how long I can sustain this level of constant reinvention without something giving. The early-2000s energy is back, sure… but so is the burnout that often came with it. Being a bit older now, with young kids at home, the pace hits differently. I don’t have the same endless energy I once did, and the late nights and constant context-switching carry a heavier weight. Finding balance is tough, but it feels more important than ever. Hopefully we can all figure out how to ride this wave more sustainably - without burning out in the process.

@rad9800 Lol didn't know there's a "Poland Street" in London 😅🇵🇱

I cannot recommend drinking lemon juice - I am currently 400ML through this 1L bottle. I look to finish the remaining 600ML over the next few hours.

@_xpn_ Cheers! I recommend using it with larger tasks, too much friction and overhead for small feats :) and yeah, constitution is great - "you must not touch anything in folder X or anywhere outside this repo" - helps keep me calm Claude won't ruin my machine (so I hope!)

@mariuszbit I started with Spec Kit last night... very nice! Gonna go through the process of using it on a few different personal projects but I like the organization and referral back to the constitution. Thanks for the recommendation!

How many people are having success with "Spec Driven Development"? Does it add anything to your workflow? github.blog/ai-and-ml/gene…

@_xpn_ Speckit all the way. It just worked for me so I stayed, never looked at OpenSpec tbh. Currently using it heavily for all my personal projects. Maybe it's just me but I don't feel like going all in vibe coding with the work code. Especially in the domain of payload dev 🙃

@mariuszbit Awesome! Do you use OpenSpec or spec kit? Do you benefit more from this at work during collaboration or is it useful in personal projects?

In collaboration with Lookout and Google (thank you 🙏) we have been working on tearing down and building detections for DarkSword - iOS exploit chain for iOS 18.4 - 18.7. Super excited for this research 🎉. Please update your iPhones. iverify.io/blog/darksword…

[RELEASE] Better late than never! Part 3 is out! Fantastic unwind information and where to find them. We went digging through .pdata, RTF Lookups, and a few ntdll internals that probably weren't meant to be touched. BYOUD dropping alongside. Enjoy 😉 klezvirus.github.io/posts/Byoud/

@kyleavery iAMSI 🙃

This seems to be initiated by OSACompile. It calls InterpreterSecurity_ScanBuffer from libInterpreterSecurity.dylib

I'm not sure when this started, but macOS seems to scan JXA scripts at runtime using the Yara rules in: /var/protected/xprotect/XProtect.bundle/Contents/Resources/XPScripts.yr

🔥 Took the Month of AI Bugs wreckage and turned it into a paper - AI Kill Chain 🧨 - Test cases and exploit chains (data exfil, rce, zombies!) - AgentHopper (a working AI virus for coding agents) 🦠 - SpAIware - Normalization of Deviance in AI zenodo.org/records/187692…

@HackingLZ Maestro + built in Cloudflare tunnel not your cup of tea?

I still haven't found a solution I love for controlling Claude cli from my phone.

Yet the offsec consulting still operates just fine. Just because there are cutting edge advanced solutions, doesn't imply they fit and apply to business/network/systems complexities and context. Offsec cannot be "solved" bc it constantly evolves, leaving models hardtime to adapt

We've seen many offsec daunting waves - KPP killing kernel threats, MOTW killing initial access, CAP killing Device Code phishing, SMB signing or leap channel binding killing ntlm relaying, CSP or JS type security killing XSSes, DEP/ASLR/Heap hardening killing Exploits

Just like ML didn't sunset Radiologists (Andrew NG, 1), retrospectively trained models will struggle to fit optimally to the fastly evolving domains like Cybersec. After all they suffer from cutoff, underfittitng, limited context and lack of taste 🙃 (1) radiologybusiness.com/topics/artific…

1,184 malicious skills were found on OpenClaw's ClawHub marketplace - stealing SSH keys, crypto wallets, browser passwords, and opening reverse shells. One attacker uploaded 677 packages alone. The #1 ranked skill had 9 vulnerabilities and was downloaded thousands of times. awesomeagents.ai/news/openclaw-…

The macOS Hardened Runtime isn’t a dead end for in-memory execution. In his latest post, @kyleavery looks at the 'allow-jit' entitlement and demonstrates shellcode execution in apps that have it. outflank.nl/blog/2026/02/1…

@HackingLZ Actually the more I use it, the more I dislike staring at iterm2/wt tabs 😅 Built in remote access via Cloudflare Tunnel, SSH support, audio notifications (jingles from myinstants.com + posh that plays random tune), tabs organization. And it works with CC skills too 🍻

@mariuszbit You still liking this? I was going to set it up last weekend hopefully I find time tomorrow.

runmaestro.ai == pass-through 1. install codex, claude, gemini 2. plug them to maestro 3. ??? 4. keep on crunching

* gemini in the workings tho

@HackingDave Probably worth also mentioning current trend for Spec-Driven Development pushed by spec-kit: github.com/github/spec-kit This kit brings a set of slash commands that define SDD pipeline made of steps such as constitution, spec, plan, analysis. Really solid thought-out approach

Obra/superpowers as a skill in Claude is a must-have imo for Claude. It forces claude to create sub agents for specific tasks, stops short cuts, keeps iterating through into a desired solution and has cut error rate on development down substantially. Super cool. Easy to add: /plugin marketplace add obra/superpowers-marketplace /plugin install superpowers@superpowers-marketplace For small things, it will take a bit longer (meaning more token usage), but for big projects - this is a night and day difference. github.com/obra/superpowe…