

Jacob Krell
3K posts

@hackerfren
Cybersec/AI expert | Hacker | Pilot | Lifter | OSCE3, CISSP, CCNP, CSIE | Top 20 Hack the Box | CTF Reviews and Writeups | meme magic 🐸





Bungulators, I have claimed AT&T and it is ours now.

GitHub just confirmed that 3,800 internal repositories were stolen… through a single VS Code extension. Not a zero-day. Not ransomware. A developer plugin. This is TeamPCP’s FIFTH supply chain compromise in ~3 months, and it highlights a massive blind spot most organizations still ignore: IDE security. Most companies heavily govern: ✅ SaaS apps ✅ Cloud infrastructure ✅ Production environments …but allow developers to install extensions with virtually unrestricted access to: ⚠️ source code ⚠️ credentials ⚠️ cloud tokens ⚠️ local systems The attack surface has officially moved upstream, into the tools used to WRITE the code. If your organization hasn’t started governing developer tooling, extension usage, and workstation trust boundaries, now is the time. The GitHub breach wasn’t the anomaly. It was the warning shot. Read @jacob krells latest research here: na2.hubs.ly/H05FnMT0 #CyberSecurity #SupplyChainSecurity #DevSecOps #VSCode #GitHub #SoftwareSecurity #ThreatIntelligence #Infosec


Cloudflare's security team spent the last few weeks testing Anthropic's Mythos against fifty of our own repositories. What we learned about offensive AI, why faster patching is the wrong reaction, and what the architecture around vulnerabilities has to look like next. cfl.re/49BRUqW



1/ We are sharing additional details regarding our investigation into unauthorized access to GitHub's internal repositories. Yesterday we detected and contained a compromise of an employee device involving a poisoned VS Code extension. We removed the malicious extension version, isolated the endpoint, and began incident response immediately.


@yacineMTB Absolutely. Having marketed bad products and good products, it is insane how much easier it is to market good products because they deliver on the promises you get to make.


The challenge is real. Agents may evade traditional detection through dynamic code generation, behavioral variance, and by operating below anomaly thresholds. Attacks will be distributed across models and vendors — there is no single vantage point to see the whole picture.








