I started GreenFlux as a freelance developer in 2016, and wrote a lot of blog posts and tutorials for Google AppSheet and Apps Script, Appsmith, FileMaker Pro, and AI and API integrations over the years.

Now I’m starting a newsletter to go with each blog post, with extra content about:

MacOS Finder lacks the folder thumbnail preview that Windows has had since XP! Let’s fix that with a little Python! 🐍

I’ve tried a ton of different ‘vibe-coding’ tools, and I keep going back to Claude Code. Anthropic’s Claude model is well regarded as the best coding model, and I’ve found the best experience is to use it directly in their own CLI tool instead of using Claude through Lovable, Cursor, or some other AI developer tool.

My normal workflow is to create a new folder and CLAUDE.md file locally, then open the terminal in that folder and run /claude to start a chat and ask if Claude has any questions before starting the project. Or, for working on an existing repo, you can download the repo locally and then open the terminal and run /init, and Claude Code will scan it and write the CLAUDE.md for you. This gives the model a starting point that is referenced in every prompt.

You can trigger Claude Code to run certain actions locally in the terminal using Hooks. I had heard about hooks and it sounded interesting but I didn’t have a use case yet. Then Claude told me “You’re absolutely right!” for the 100th time! when I corrected it on a stupid error, and I thought 🤔:

So I made a counter! ☑︎

Recently Kevin Blanco (Sr DevRel @ Appsmith) and I went to the Shift Miami Conference to learn how others are leveraging AI agents in their daily work. We took a break from the conference to have a real conversation about agent adoption in production environments and the community response has been amazing (3.2k views, 650+ likes)!

Lately I’ve been learning a lot about knowledge graphs, and how they offer some significant benefits over vector databases for RAG (retrieval-augmented generation). And one of the main sources of content I have found is Jason Koo from Neo4j.

If you’re not familiar with GraphRAG, it’s a form or retrieval-augmented generation that uses a graph database instead of a vector database. There are some significant benefits to using graph databases that enable queries and reporting that would be impossible with SQL or vector databases. And Jason is the best source of content on adding graph databases to any tech stack.

Here’s his latest video on using the Neo4J MCP

If you enjoy this newsletter, please share it with others in your network. Got a question or comment? Login at news.greenflux.us to share your thoughts.