We've written about Erebus getting closer. Today it's out: an AI-native C2 framework, architected from the ground up for AI-assisted operations. Human-in-the-loop by design, with secure infrastructure and a modular, cross-platform core.
It's solo-built, and it's been validated against non-trivial HTB engagements before this release. That's the bar we held it to: not a demo that works once, a framework that holds up under real operator use.
$ git clone https://github.com/KKingZero/Erebus.git
status: open source, out now
Why AI-native, not AI-bolted-on
Most C2 frameworks were built for human operators first, then had automation layered on top. Erebus flips that: the AI reads state, plans next steps, and prepares tasks as the primary control loop. The operator isn't driving every click. They're the approval gate on anything sensitive.
That's not unattended exploitation. It's supervised speed: AI control with human judgment where it matters.
Why open source
A C2 framework touches the most sensitive part of an engagement. If you can't audit it, you're trusting it blind. Erebus is fully open: the teamserver, the implant, the operator CLI, and the AI agent layer are all there to read, build, and verify yourself.
Get it
Erebus is for authorized red team engagements, internal security testing, and research environments with explicit permission. Clone it, read the code, run it against your own lab or engagement scope.