7 Claude Code workflows I use every day (most people know zero)

I've been running Claude Code as my primary coding tool for months now.

Most people install it, type a prompt, and hope for the best.

Here's what happened when I stopped doing that.

I found 7 workflows that changed how I build with AI. None of them are obvious from the docs.

1. Start every session in plan mode

Claude Code kept going down time-consuming, token-burning tangents that took hours to unravel. Now I start every session in plan mode. It presents a markdown doc of exactly what it's going to do. I review it before sending the agent to work.

2. Run /init on every new project

This creates a CLAUDE.md file that you (or the AI) can populate with project instructions. Claude reads this file automatically every time you start a new session. No more re-explaining your codebase. No more tangents.

3. Save repeatable workflows as skills

If you find yourself doing the same thing over and over, ask Claude Code to save it as a skill. Then invoke the entire workflow with a single command. I use this for committing and pushing to git.

4. Run /insights weekly

This might be my favourite command. It reviews all your sessions from the last 7 days and generates an HTML report. What you worked on. Where you wasted time. Mistakes you kept making.

Mine told me a single 6-hour session is far less effective than starting a new chat for each job. It even suggested updates to my CLAUDE.md file and spotted automation opportunities.

5. Open multiple instances

I run 4-5 Claude Code windows in Warp (my terminal app). One window per job. I give each window a name and tab between them to check progress. Each agent stays focused because its context is clean.

6. Use sub-agents for complex projects

Claude Opus 4.6 can spawn specialist agents that work on your project simultaneously. One for UX, one for copy, one for debugging, one for frontend QA. The master agent analyses all their feedback and gets to work. It's like hiring a senior dev with a team. Costs more tokens, but the output quality jumps.

7. Teleport your session

I usually run Claude Code on my MacBook, but when I'm on the go I hit teleport. It sends the coding session to the web version. I continue on my phone, then teleport back to my Mac later. Same context. No lost work.

The pattern across all 7: Claude Code rewards structure over speed.

Give it a plan before you give it permission. Give it context before you give it code. Give it constraints before you give it freedom.

Most people treat AI coding tools like a magic wand. The ones getting results treat them like a junior dev who needs clear instructions.

Want to learn the exact prompts and frameworks I use to build with AI? Check out my Prompt Writing Studio course