If Moscow rules meant watch your back, London rules meant cover your arse. Slow Horses

Mini stack for a secure playground

Following up on my post regarding power users and PATs. As discussed then, we can assume the following: Power users will automate. You only get to choose whether they do so within your security posture or against it.

I choose the former because there are more of them than there are of me and they've been hired exactly for that kind of mindset. Hence: they need something safe to automate against and somewhere safe to build. Here's the substrate I've been building:

Why? Because that way I can focus dev efforts on the functional value add and more easily update authorization logic without having to do a full redeploy. Traceability, non repudiation and easy contextualization when digging through the opentelemetry data to pinpoint issues.

How? one Nix flake that defines a standard QEMU VM to program against with:

Output:

And of course it's very fun to play with and run fire drills against!