2. GLM-5.3 and OpenCode for cybersecurity

3. Primarily custom built harness. Rust centric, heavy on the production engineering gates and a bunch of graph networks for context control.

4. Custom.

5. No.

My general ethos is that if taste is what is important going forward, my code generation must have it's own taste. If you can download my tooling off the shelf, I am replaceable by anyone who can do the same. My end goal is to drop Claude Code once the harness prices start to drift to API pricing. In the meantime, anything that can be automated into traditional and efficient code will be one more thing in my control regardless of future token prices. When I hit my limits for the week, I spend the remainder of the waiting time going over the framework data to see what are my largest sources of consumption and iterate in automating that more.

I was unhappy with everything out there - not secure enough, not flexible enough, not composable enough, not nix enough. The whole concept how we use AI is garbage, having 10 terminals open with agents working in different branches. I build workflows, very complex, multi model, multi harness, parameterized - nix defined workflows that run in highly isolated, fully composed environments. I have a few dozen workflows already, in vx-maintainer: quick-fix, change-management, spec-implement, red-green, sync-branches, ....

Each of those workflows does something, and this extremely well. Every time a workflow fails, the workflow gets some massage to make it more resilient, more flexible, more correct. Workflows are usually design that restart just recovers the last state and continues work. My land-feature workflow has at least 40 steps, most of them deterministic scripts. For example, if rebase is true, a mechanical rebase is started, if this fails, an agent rebase agent is started. Full multi panel + review tool pipeline. With time, the workflows start to be super sophisticated, for example: review is cached, but the review is pinned to the commit hash. When restarted and rebase is on, a ast-diff is generated to check if the review can be reused. Stuff, no normal person would do :)

vx-maintainer has an own bug / change tracker. Normal bugtrackers are build for humans, such agentic system need a tracker that works for both.

Then I have a special agent, the queue-manager, his goal is to keep the workflow engine busy. He queries the vx-meta tracker for changes and workflows to get implemented.

Workflows are semi-automatic. You can have tty and ui programs running inside the workflows. You have decision steps and agents can ask the operator.

And this is just the surface of it.

I never was able to get so much stuff done and had so much fun getting a project build. And fucking safe - no more fear when updating dependencies.

1. subscription: claude max 20x

2. other subscription: nope

3. harness: vs code with claude plugin

4. orchestration: no

5. broader agents: grok for outreach.

my development setup is a high-end local esxi server (multiple internet cons and a graphics card). The server has about 30 different OSes on it (static ips / ssh setups for all vms.) I use a VM on that machine to run my IDE with full access to the other VMs. I have good snapshots for all the VMs so if anything breaks they are easily restored. This setup makes it easy to develop software for any OS and also test complex software interactions.

For hosted internet services: I use a single OVH debian machine. I pre-pay everything over a large period so the bills are always paid. That keeps it reliable as I can say that even if I go broke in the future I'll have paid it for quite a while (useful in this economy, lol.)

Session notes: I feel like the best approach is for us all to just maintain a personal wiki for our projects. But I must admit I've been lazy in that respect. All my attempts to manage my notes never works because im such a messy, chaotic person. My natural state is disorder. So I'll just (1) destroy it by accident (2) forget to back it up (3) forget to use it (4) get tired of using it. I envy those super organised people.

Backup: my services are backed up locally. My local repos are backed up to github, email, and remotely. Agents deleting anything by accident isn't something i have to worry about. They have ruined VMs before. I can easily restore them. So far I've only ever lost notes because of laziness.

Security: my security setup could be vastly improved by using discrete key pairs with restricted permissions for github. What I worry about most is just things like prompt injection and the agent accidentally downloading malware. Lateral traversal out of vms is hard but not impossible. I'd wager there's still lots of private stuff on the vms too. I monitor everything closely at least and can restore a snapshot if i get paranoid.

2. OpenRouter PAYG - glm or opus for most things

3. OMP or CC

4. Zellij/Vim — some things never change

5. No

I avoided building anything custom until I had to, and I avoided adding more than a single agent until I had to.

My preferred way to use AI for coding has been to work closely with a single coding agent with the code open. That changed wth 5.6-Sol and a handful of skill files. I stopped needing to watch closely to keep it from fucking up the codebase. With the right request, that combination will get the job done 9/10 times.

That unlocked a new workflow: as soon as a change is requested, a coding agent reads it and gives me a PR. I want a complete proposed code change in front of my eyeballs by the time I open the change request myself.

So I have a custom solution in place to manage a handful of agents aligned with this workflow.

Custom solutions are so cheap and easy now... it's really hard to put up with solutions that are only mostly right for the job.

So I made my own. Including patching onnx myself to run parakeet model (way better than whisper)

Fully FOSS: voiceio.dev

2) I tried a local Qwen model but my poor old 1660 Ti did not enjoy it, and it wasn't all that useful.

3) I primarily use OpenCode running in a podman container for actually working on stuff. Also use the default free chats with Gemini 3.5 Flash for general searching, ChatGPT for brainstorming, and Claude to answer more code related things if they come up.

4) Nope.

5) Not at the moment. I've been eyeing using Hermes, but since I don't feel comfortable paying much at the moment I haven't really tried.