It's not in my AGENTS MD but it's literally my most used piece of text by a mile. "Issue happened at 10:49-50. Check logs, if we don't have logs add logs to understand the root cause, then rebuild and repro with computer use. if we do have logs and you're sure about the root cause then explain the root cause, and give me the best options to resolve this issue." This puts me in the driver's seat and doesn't let the agent start implementing until I discuss with it and I'm sure that it understands the real issue. When I don't do this it keeps trying to solve things in an opaque way and without making 100% sure that it's solving the right thing. I'm sure I can add this to my AGENTS.md somehow but it would be worse than always pasting this or saying a variation of it in wispr flow. Have a similar one for features implementation too. I keep discussing till the cows come home and even tell it to ask GPT 5.5 Pro until we have a very clear picture (no plan mode ever). Then I implement with a /goal (used to use beads/subagents but they're not needed anymore with 5.5 Codex CLI)

ghostex.dev supports Fable 6 and GPT 6!

Jeff Bezos reveals the simple phrase that saved him countless arguments running Amazon "Disagree and commit is a really important principle that saves a lot of arguing" "One of my direct reports would want to do something. I'd think it was a bad idea. We'd go back and forth and I'd often say, you know what, I don't think you're right, but I'm going to gamble with you" "You're closer to the ground truth than I am. I've known you for 20 years, you have great judgment" "At least then you've made a decision and I'm agreeing to commit to that decision. I'm not going to be second guessing it, sniping at it, or saying I told you so" "I'm going to try actively to help make sure it works. That's a really important teammate behavior"

1/ This week our team at @trycua worked hard to ship cua-driver v0.5. The headline is a new primitive, Agent Cursors: many background agents can now each own a cursor, scoped to a non-overlapping action space.

It's not in my AGENTS MD but it's literally my most used piece of text by a mile. "Issue happened at 10:49-50. Check logs, if we don't have logs add logs to understand the root cause, then rebuild and repro with computer use. if we do have logs and you're sure about the root cause then explain the root cause, and give me the best options to resolve this issue." This puts me in the driver's seat and doesn't let the agent start implementing until I discuss with it and I'm sure that it understands the real issue. When I don't do this it keeps trying to solve things in an opaque way and without making 100% sure that it's solving the right thing. I'm sure I can add this to my AGENTS.md somehow but it would be worse than always pasting this or saying a variation of it in wispr flow. Have a similar one for features implementation too. I keep discussing till the cows come home and even tell it to ask GPT 5.5 Pro until we have a very clear picture (no plan mode ever). Then I implement with a /goal (used to use beads/subagents but they're not needed anymore with 5.5 Codex CLI)

i prefer apps personally (codex app) but dude this is epic, this would've been a life saver for the half a year we were living in terminals haha

I really tried to use it but it gets confusing when you have a bunch of threads going in a few projects. Made ghostex.dev to solve the UX papercuts I faced using cmux for agentic dev (also uses Ghostty terminals). Please let me know what you think 🙏🏻 Just released an Android app to connect to running sessions on my macbook live (setting to make all sessions use zmx for persistence and attach/detach flow)

Me with your feedback :) Already have a great base with @ghostex_dev, would love your opinion about what to focus on next. The biggest differentiator is the app uses native Ghostty terminals so it uses way less resources and has the best compatibility with all Agent CLIs. T3 Code is integrated and works well for people who prefer GUI too. Also includes a lite VS Code embed for code and MD files editing. Recently added a Rich Prompt Editor to edit your Agent CLI prompts. Really proud of this one, no one else has it working so well. Please let me know if you ever get the chance to try it, I'm not great at marketing but the app's UX and tech base is really one of the best you'll find (not perfect yet but will be soon).

I know you saw this before but I can't wait to hear your thoughts so self promoting again 😅 Presenting @zmux_app: My temu Codex app that lets you place native Ghostty terminals any which way you want. Cooking up a huge update that will come out soon with: - A rich editor for CLI prompts 🤯 (ctrl g and then edit with all your hotkeys in a floating window, even image pasting is supported) - Pop out mode - Tabs and splits - Better hotkeys - Lots of bug fixes Testing droid is top of my list for this week too.

Even better is to totally abstract bash away, and block requests using an intelligent system when they match specific disallowed patterns. Can work like how ad blockers/antivirus that recieves new definitions every once in a while work. I think there's good business opportunity here for selling to enterprise if you can also insure them (like Cloud Strike and similar companies)

More and more people are asking me about testing resources so let's put everything I've written in one post. Bookmark, share, and, most importantly, please read these. The True Purpose of Testing epicweb.dev/the-true-purpose… Developers often overlook the fundamentals and rush into writing tests without properly understanding what a test is and what is its function. No test is inherently useful just because it exists. Read this one to learn what makes it useful. The Golden Rule of Assertions epicweb.dev/the-golden-rule-… There's a lot of debate over what makes a good test. In this one, I'm defining a short and objective way to grade a test's quality no matter the language or the tested system. This is, without a tinge of exaggeration, a game-changer in how you approach your tests. Anatomy of a Test epicweb.dev/anatomy-of-a-tes… Let's talk about the building blocks that make up any automated test. From JavaScript to Go and Rust—these blocks power tests everywhere. Know your blocks. Implicit Assertions epicweb.dev/implicit-asserti… Did you know there's a way to express expectations in tests without writing "expect"? Those are called implicit assertions and they are tremendously powerful because they help you express more by writing less. Inverse Assertions epicweb.dev/inverse-assertio… Sometimes you need to assert that something did not happen. That can be tricky, especially if that something is asynchronous. The last thing you want are false positives. What you actually want is inverse assertions. Making Use of Code Coverage epicweb.dev/making-use-of-co… Code coverage has been an ongoing debate in the engineering circles. Is 100% code coverage in tests good? Bad? When should you strive for it? Why do people say it's harmful? I'm answering all those questions in this one and giving you practical tips on when to use (and not to use) code coverage. Good Code, Testable Code. epicweb.dev/good-code-testab… You've gathered by now that some code is easier to test than the other. But why? Let's take a look at the characteristic of code's testability, what defines it, what is its relationship with complexity, and how to make your code more testable. What is a Test Boundary? epicweb.dev/what-is-a-test-b… Automated tests rarely involve your entire system (yes, even the end-to-end ones have exceptions). There's often a place where you draw the line. The boundary. Learn what it is and how to use test boundaries efficiently to focus on the exact behaviors you want to test. Be S.M.A.R.T. About Flaky Tests epicweb.dev/be-smart-about-f… Flakiness is the scourge of reliability. If you've written a test before, you likely had experience with flakiness. But what is it at its core and what causes it? And how should you deal with flakiness? Writing Tests That Fail epicweb.dev/writing-tests-th… You write tests for them to fail. We all enjoy a green CI, but the true value of tests is when they fail. What matters is when and how they fail.

Forgot to mention: • Built in zmx/tmux/zellij support (can continue via ssh then use zmux cli to attach) - beta but working well atm, especially zmx. • Automations and cross agent messages coming very soon • Better worktrees support coming very soon. Want to nail UX

But Ghostty with Codex App UX and full Pi support is even better! 😤 github.com/maddada/zmux • Native macOS app • Uses native Ghostty terminals (best cpu/ram use) • Inspired by Codex App's UX • Auto sleep unused terminals • Auto resume prev sessions • CEF embedded browser w devtools & profiles • Light embedded VS Code based editor & git manager (gh PR extension works) • Embedded T3code • Agent CLI integrations • Rich prompt editor (in testing) Much more stuff. Please star and let me know what you think 🙏🏻 Been working on this for months now.

If you have memories enabled on Codex CLI -> Expirements. Turn them off right away! Enabled it for few hours to test and then checked what it wrote and it was 6000 lines of the most useless garbage ever. Make sure you have nothing in this folder & file: /Users/madda/.codex/memories/MEMORY.md

Me irl

Yep, tmux didn't work well for me. I created a tool to bring the best flow into VScode. Supports all the main agent CLIs and allows using GUI through T3code (dpcode fork) Has auto naming, search, actions, notifications, floating bar, integrated browser, up to 3x3 split, MIT, sleep terminals, persistence (w/ xterm.js & ghostty), can connect from phone, and is very customizable. Please tell me what you think 🙏🏻

I'm a very long time windows power user and a recent convert to macOS (Got an m2 max ~1 year ago cuz f laptop battery life with x86) I really didn't expect I'd be saying this but you really can't convince me to go back to windows. macOS is leaps and bounds better once you customize it to fit the way you work. And now with everything AI coming out and supporting macOS first it's even more of a no brainer. Worse thing about macOS is window management and finder but macOS is insanely customizable with open-source apps (many of which I'm contributing to or forking for myself). I REALLY hated apple before. And I will never switch to iOS, it's so locked down I don't feel like I own the device, but macOS is like very linuxy if you know the right apps to download. I honestly believe that unless Microsoft throws powershell in the garbage and give us native unix then they're going to be dead for app/web dev in a few years. Everything is coming out for macOS & linux first. My macbook works almost exactly like my windows machine now, even down to the hotkeys, alt tabbing, the dock, mouse, and it's been SUPER polished since people on macOS side make money on these utility apps and the quality of apps is higher overall. Please let me know if interested I can write a guide, been thinking about doing it for a while. I don't gain anything from shilling for Apple, still dislike the company, but what they have with macOS M processors is the best of all the current options hands down. All the best.