Anthropic slipped Cowork’s most interesting behavior into a support article. I turned it into a Karpathy-inspired system that gets smarter without writing a line of code. A read for builders, PMs, and anyone who refuses to ship without thinking.
TL;DR Cowork Self-Improving Loop = Karpathy Auto-Research pattern + recurring tasks.
Somehow, Anthropic managed to hide Cowork’s best kept secret in a support article about recurring tasks.
Very few people seem to know about it, and I haven’t seen anyone intentionally building on top of it.
This guide is my attempt to change that.
By the end, you’ll be running a self-improving loop I built on top of Karpathy’s Auto-Research Pattern, adapted for Cowork.
So even if Claude Code feels intimidating, you’re covered.
I’m an AI Product Manager and builder. I write Product with Attitude, a newsletter about building with AI and developing critical AI literacy through practice. If you’re new here, welcome! Here’s what you might have missed: → Claude Cowork Guide for Power Users (2026) → Claude Skills Are Taking the AI Community by Storm
Join 17K readers and learn AI the only way it sticks: through immersion in real experiments and real projects.
- How to turn Anthropic’s Claude Cowork into a self-improving AI agent
- How Cowork already rewrites its own scheduled task prompts after the first run, even though almost nobody has noticed.
- The full no-code architecture: context.md, folder instructions, and the improvement directive that makes the loop work.
- Step-by-step tutorial from project creation to a running automation loop.
- Ready-to-use prompts.
- Risks, guardrails, and what happens after 20-30 runs.
A workflow could sit at half its potential for months and we’d never know because we stopped looking.
Every AI workflow decays.
We set up an AI workflow. It works beautifully at first.
Then the world around it changes: files move, naming conventions change, things get re-organized, new edge cases pile up.
The instructions stay frozen, our needs don’t.
The conventional fix would be to open the prompt/Skill, read through it, edit it, test it again. Rinse, repeat, lose a chunk of our busy week to maintenance.
And it has a blind spot: we only improve what we've already seen go wrong.
New inputs, new edge cases, new failure modes, those slip through.
A self-improving loop changes that. It makes improvement part of the task, not something you remember to do when things go wrong.
Want to read the rest? The full post is here → Read on Substack
- (Karo Zieminski, authored, "I Built a Claude Cowork Loop That Improves Itself. Here's the Exact Setup.")
- (Product with Attitude, published, "I Built a Claude Cowork Loop That Improves Itself. Here's the Exact Setup.")
- (This guide, is, my attempt to change that)
- Anthropic, Attitude, Auto, Claude, Claude Code, Claude Cowork, Claude Cowork Guide, Claude Skills Are Taking, Community, Cowork, Cowork Self, How Cowork, Improving Loop, Improving Loops Fix, Inside How
- AI agents, AI product management, Karo Zieminski, Product with Attitude, Substack, critical AI literacy
#ProductThinking #AIForProductManagers #ProductStrategy #Vibecoding #AIAssistedCoding