MCP Permissions in Teams: The Problem Nobody Solved Yet
When everyone on your team can create PRs on any repo or trigger PagerDuty, you've got a problem.
When everyone on your team can create PRs on any repo or trigger PagerDuty, you've got a problem.
Non-engineers are submitting PRs now. Here's how to keep your codebase safe without killing the momentum.
AI tools make everyone faster. But when the whole team speeds up, some interesting problems emerge.
You shipped it fast. It worked. Now it's broken and you have no idea why.
After months of daily use, I've learned to recognize when Claude Code is about to waste my time.
How I actually use Claude Code when something's broken in production. Real workflow, not theory.
I kept typing the same prompts over and over. Slash commands fixed that.
A handful of things that made my Claude Code workflow way faster. Nothing groundbreaking, just useful.
Your project memory file is the foundation of every Claude Code session. Here's how to write one that actually helps.
Hooks can feed Claude information automatically. Here's how to stop the repetitive context questions.
Hooks let you run shell commands at key moments in Claude Code's lifecycle. Here's what they are and why you'd want them.
I kept over-engineering my sub-agent setups. Here's what I learned about finding the right boundaries.
A hands-on tutorial for building your first Claude Code sub-agent. We'll create a code reviewer that checks for security issues and best practices.
The beginning of my journey documenting daily AI interactions and discoveries.