"When you externalize parts of session logs to a shared space where people and their agents can collaborate, multiple benefits accrue. For people it provides transparency and accountability. Decisions and tactics aren’t squirreled away in dot file on a per-machine-per-user basis. They are accessible to the whole team both interactively and by means of GitHub APIs that were formerly daunting but now easily wielded by agents on our behalf.

For agents, GitHub is a place to record context, drawn from current work, that powerfully informs future work — again by way of APIs that agents easily wield."

https://blog.jonudell.net/2026/06/02/how-to-make-best-use-of-git-and-github-for-ai-assisted-software-development/

#git #github** **#codex** **#claude** **#bram** **#xmlui #PracticalInternetGroupware

This is the latest release of a new desktop app (Bram) that helps you make best use of git and GitHub for AI-assisted software development:

https://github.com/judell/bram/releases/tag/v0.1.35

That's for today. The release note tells a story that is enabled by LLMs (obviously) but also by the git/GitHub workflow that Bram encourages and supports.

I am not actively soliciting more beta testers, the current handful are pushing things along at just the right pace. But if this sounds intriguing, by all means take a look. The idea is to make AI-assisted dev work safe, orderly, and accountable. It's working pretty well for us so far.

#bram #git #github #claude #codex #xmlui

XMLUI development with Claude Code or Codex runs on a two-stroke engine. The agent uses the MCP server to find out how to add a feature to your app. Then it reads the app's semantic trace to find out how well it worked, or what to fix if it didn't.

If you want to put yourself in the driver's seat for a test drive, it'll take 5 minutes.

https://www.xmlui.org/blog/xmlui-for-llms

#xmlui #mcp #observability

XMLUI's Markdown component supports codefences that can define complete apps: a Main, one or more user-defined components, and an API. You can express quite a lot of interesting behavior in a few dozen lines of code written between triple backtics. I've been really excited about this capability but I had to write this post to fully articulate why: it's all about reproducibility.

https://blog.xmlui.org/blog/xmlui-playground

#XMLUI #Documentation #ReproducibleSoftware

Here's how to make an XMLUI table responsive: https://docs.xmlui.org/howto/make-a-table-responsive

I just added the example to the Howto section in the docs. Now when I ask an agent how to do this, the XMLUI MCP server should guide the agent to the working example and cite the URL.

I tested.

It does.

This is what I mean by testable documentation (https://thenewstack.io/how-to-upgrade-your-mcp-server-with-context-engineering/).

The agent is anchored to ground truth: a working example in a live playground.

#XMLUI #MCP