I've tried every task management system. Jira, Linear, Notion, plain text files, sticky notes on my monitor. They all have the same problem: you have to maintain them. You have to context-switch out of whatever you're doing, open some app, update some ticket, move some card across some board. It's friction, and friction means things don't get logged.
A few weeks ago I started using Claude Code as my task manager. Not a purpose-built app. Not some fancy integration. Just Claude Code, a git repo, and markdown files. It's the best system I've ever used, and I want to explain why.
The setup
The whole thing took about two minutes. I told Claude Code to create a new repo with a simple structure: one markdown file per day, labeled by date. Each file has two sections, tasks completed and tasks remaining. That's it.
tasks/
2026-03-03.md
2026-03-04.md
2026-03-05.md
2026-03-06.md
2026-03-07.md
Each daily file looks something like this:
# 2026-03-07
## Completed
- Refactored auth middleware to handle token refresh
- Fixed flaky integration test in CI
- Code review on Sarah's PR for the new dashboard
## TODO
- Write migration script for user preferences table
- Follow up with infra team about staging deploy
- Draft RFC for new caching layer
## Notes
- Caching RFC needs input from the platform team before next Tuesday
- Migration script depends on schema changes landing first
Nothing revolutionary. The magic is in what happens next.
Standup in 10 seconds
In the repo's CLAUDE.md file (Claude Code's project instructions), I added a simple directive: when I ask for a standup update, read today's file and yesterday's file, then give me a concise summary of what I did yesterday, what I'm doing today, and any blockers.
That's the entire prompt. No fancy engineering required.
Now when standup hits, I just open my terminal and ask Claude for my update. It reads the logs, pulls out the important stuff, and gives me a clean three-liner. No scrambling to remember what I did yesterday. No scrolling through Slack messages trying to reconstruct my afternoon. It's all there because I logged it as I went.
The key insight is that logging tasks in this system is low-friction enough that I actually do it. When I finish something, I type a quick line. When something new comes up, I add it to the TODO section. It takes five seconds because I'm already in my terminal. There's no app to switch to, no board to find, no ticket to update.
Throughout the day
The workflow is dead simple. I keep the daily file open (or just tell Claude to add things) as I work. Finished a PR? Add it to completed. Got pulled into a meeting about a new project? Add the follow-ups to TODO. Realized something is blocked? Add a note.
I also use it for things that aren't today's problem. If someone mentions something that needs to happen next week, I'll note it with a date. Claude can scan across files, so nothing falls through the cracks. It's like having a second brain that actually remembers things, which is useful when you're juggling multiple workstreams and your actual brain is full of implementation details.
The real payoff: year-end reviews
Here's where this system goes from "nice productivity hack" to genuinely career-changing.
At my last job, every quarter we had to write self-assessments. Every year we had to compile those into an annual review. And every single time, I'd sit there staring at a blank doc trying to remember what I did four months ago. I'd dig through old PRs, scroll through Slack, ask teammates what projects we shipped together. It was awful and I always felt like I was underselling myself because I simply couldn't remember half of what I'd done.
With daily markdown logs, that problem disappears completely. When review season comes around, you point Claude at the entire tasks directory and say "summarize my work for Q1." It reads every single daily file, identifies the major projects, pulls out the themes, and drafts a comprehensive summary. Nothing gets missed because everything was logged in real time.
This isn't just convenient. It changes the quality of your reviews. Instead of vague statements like "worked on improving system reliability," you get specific, dated accomplishments: "Refactored the auth middleware on March 7th, reducing token refresh failures by 40%." That level of detail makes you look organized, thorough, and on top of your work, because you are.
Why this works better than everything else
The reason every other task management system failed for me is that they all require you to do something unnatural. Open a separate app. Navigate to the right project. Find the right ticket. Update the status. Add a comment. That's five steps between "I finished this thing" and "it's logged."
With Claude Code and markdown files, the gap between doing work and logging work is basically zero. You're already in your terminal. You're already talking to Claude. You just say "add X to completed" and it's done. Or you open the file and type a line. Either way, it's one step.
The other advantage is that markdown files in a git repo give you a fully audited history for free. Every change is committed. You can see exactly when you added a task, when you completed it, how your priorities shifted throughout the day. Try getting that out of a Kanban board.
Getting started
If you want to try this, here's what I'd suggest:
Open Claude Code and tell it to create a task tracking repo. Tell it you want daily markdown files with completed and TODO sections. Add a note in CLAUDE.md about generating standup summaries. Start logging your work.
That's it. No configuration. No integrations. No learning curve. Just start writing down what you do every day in the simplest format possible, and let Claude handle the synthesis.
It won't replace Jira for team-level project management. That's not the point. The point is having a personal system that's so low-friction you actually use it. And when review season rolls around six months from now, you'll thank yourself for every line you logged.