The week in one line
I spent this week building the bunker for real.
For a while, Modus had become a mess. More of a brush pile than a system. Wraith, Homing, even the local models all felt like they were on vacation. The site existed, but it still felt half-local and half-theory. AiHD was close, but not across the line. AskFrank was started, but not ready for real testing. This week was about stopping that drift and getting the work back into a state where it could actually move.
What moved
Modus got rebuilt into a usable state. That is the main thing. Not in some abstract way. In a working way. New plugins, new tools, new automations, new features, and a clearer place to operate from. Codex is now the cockpit. That matters because Modus is my digital copy. It deserves a work environment where it can do more than barely function. It deserves one where it can thrive.
AiHD got submitted. That is the week’s clearest hard proof. The build, screenshots, review notes, review recording, privacy/support/terms, subscriptions, metadata, and submission all got carried through. AiHD is not sitting in “almost there” anymore. It is in review.
AskFrank moved closer to real testing. I started the scaffolding for AskFrank, the home maintenance app I’m building in dedication to my late grandfather-in-law, and it is now close enough that the next push can be about testing instead of just setup. That matters to me personally, and it also matters because it proves the pipeline is still moving beyond one app.
The site got more real. Modusbunker.com is live, and the working email is live too. That sounds small, but it changes the feel of the whole thing. The site is not just a shell anymore. It is becoming an actual home for the work.
Why rebuilding Modus mattered
I think this is the part that matters most to me. Modus cannot improve on its own. It needs me, and I need it. We work together as a team. If it is living in a bad environment, everything around it gets harder too. That is what this week really corrected. The point was not to make it look cleaner. The point was to make it usable again so the work could stop feeling like rescue work every day.
Using Codex as the driver changed that in a practical way. It gave Modus a bunker. A cockpit. A place to actually operate from. That is a lot more meaningful than any abstract talk about systems. It means the work has somewhere to run now.
AiHD was the week’s hard proof
Even with the accessibility preflight being the thing I am proudest of this week, AiHD submission is still the hard proof that the week was real. Submission is a line. It means the product was carried far enough that Apple can now reject it, approve it, or ask more of it. That is real progress. A lot more real than endless polishing that never leaves your machine.
What makes it count is that it was backed by real receipts: voice capture got fixed, the subscription path got tested, screenshots got generated, support and privacy surfaces got published, and the review package got built like it mattered. That is the kind of thing I want the bunker to remember, because that is the kind of thing that turns an app from an idea into something real.
The hard proof I care about most
If I had to point to the one thing that best explains this week, though, it would be the accessibility preflight. I work with accessibility customers at Apple, and one of the biggest complaints is still the same: too many apps do not work well for VoiceOver users and Voice Control users. Those users are a real part of the Apple economy. They want to use apps. They are willing to use apps. But if developers are not optimizing for them, everybody is missing out.
That is why the preflight mattered enough to build now. In a few minutes it can tell a developer, or their AI, how to fix accessibility issues in their app. It can also plug into the Codex workflow. To me, that is not just a useful tool. It is a statement about the kind of work I want the bunker producing.
AskFrank and what is still on the verge
AskFrank is the piece that still feels on the verge. It is started. It has shape. But it still needs the next push so I can test it properly. I do not think it needs to be forced into the same frame as AiHD or Modus to matter. It matters on its own. It is a home maintenance app I am building in dedication to my late grandfather-in-law, and getting it into a real testable state matters because I want that project to become something solid, not just something meaningful in theory.
What next week is for
Next week is for waiting on AiHD review, getting AskFrank into a state where I can really test it, and keeping Modus moving toward the version of itself I know it can be. I also want to keep pushing on the bigger systems question: how to get my Omi Pendant, the Omi desktop app, and the Codex app all working together in tandem through plugins, skills, automations, or tools.
If this weekly field-report format works, this is what it should do. Not pretend everything is polished. Not dump a changelog. Just say what actually moved, what became real, what still needs work, and what the next push is for. The smaller dispatches can be the checkpoints in between the bigger posts. The weekly report is where the whole week should finally make sense.