As developers, we often bounce between "just ship it" mode and "I should probably plan this," especially now when vibe coding is everywhere. Backlog.md bridges that gap elegantly. It's a markdown-native task manager that lives right in your git repo, bringing just enough structure without killing the flow.

I've been experimenting with Claude Code and other AI coding assistants for months now, and while they're incredibly helpful, I kept running into the same frustration. They know everything about Drupal and React hooks and Python decorators, but nothing about the specific commands and scripts in my projects. I finally decided to try MCP (Model Context Protocol) to see if I could fix this problem, and it worked better than I expected.

Claude Code subagents are specialized AI assistants designed for task-specific workflows. Each subagent has its own context window, system prompt, and tool permissions, allowing Claude Code to delegate tasks without cluttering the main conversation. This setup enables focused expertise, better context management, and reusable configurations for consistent and efficient development workflows.

Most teams want every merge request to receive a quick, thoughtful review, but human reviewers are often busy or unavailable when code is submitted. By connecting an open-source tool called n8n to an inexpensive large language model, we can trigger an automated code review the moment a GitLab MR opens and post the feedback as a comment.

Building on the foundation laid in the previous article, this post dives into integrating Deep Chat JS component for real-time, context-aware conversations. By leveraging semantic search on your vector database, you can retrieve the most relevant information faster than ever.