Recent hype about loop engineering made me realize that im missing one thing, and that is an agent that should be doing my job at the end of my loop: real user test

i developing a huge b2b ecommerce project codebase. and for every new feature or fixes my job is always to test it on the browser. yuck.

so i have made some research and there is TestSprite, but its not what i really want, i cant even self host it.

so i built OpenUser. now i just engineer the loop with simple prompt: complete {feature} and use /openuser to test it.

it store user persona, struggle, checkpoints, console logs, network logs. the agent has access to all and just fix everything.

this is the real replacement of me at the end of the loop.

everything runs locally, with any coding agents and model.

try it: https://github.com/manalkaff/openuser npx openuser-cli

I built a UI framework named Domphy because I could not learn React although I tried many times. I mean I could not understand React code which is hidden behind multiple layers, especially when some of them use React ecosystem libs. Maybe I am an architect turned developer, so I had no foundation of programming, or I expected something clear and clean.

I just think that UI should be simple — HTML and JS are enough and just need some way to make it stateful instead of making it more complicated. My idea: just using JS objects to reflect HTML, and using functions for state. When I need to reuse a component I had one more concept — just make a partial (I call it a Patch) to add props to the main object (but native props still win). With a component-based approach you get deep nesting and exploding props, but with patches you don't. The example below:

  import { ElementNode, toState } from "@domphy/core";
  import { tooltip } from "@domphy/ui";

  const count = toState(0);

  const App = {
    div: [
      { h3: (listener) => `Count: ${count.get(listener)}` },
      {
        button: "Increment",
        onClick: () => count.set(count.get() + 1),
        $: [tooltip({ content: "Add one to the count" })],
      },
    ],
    style: { display: "flex", gap: "8px", alignItems: "center" },
  };

  const root = new ElementNode(App);
  root.render(document.getElementById("app")!);

Right now I am the only one using Domphy, for around a year, for creating SketchUp and Revit plugins in the AEC (Architecture, Engineering, Construction) industry. I created Domphy before AI code generation took off, to make code humans can read and understand clearly, but now AI can build UIs with React well, so sometimes I feel my work is meaningless. But I still use Domphy for my apps, because I feel more confident when I need to read and edit UI code when the AI gets stuck.

I am a hardware engineer and security researcher and I've been wondering whether my work could be partially automated, so I can focus on other topics as well, so I build revenant - a LLM powered (Claude, OpenAI, local AI) toolkit that builds around radare2, ghidra etc and can fully automatically analyze firmware, implement open source skeletons incl. pinouts, hardware bringup, peripheral bringup etc. or can even 1:1 replicate existing firmware so old hardware can be resurrected with modern toolchains.

Some applications are:

- Give old hardware new life

- Security Analysis of shady firmware

Check it at: https://github.com/DatanoiseTV/revenant

I would love some input on this and maybe some recommendations for improvements.