Brooke Wright
Brooke Wright · @wright_mode
FREE RESOURCE

Make AI actually finish the job

The unlazy skill breaks every task into checklists with a line defining what finished looks like — plus the parallel-agents prompt so it stays fast.

The repo link 2 copy-paste prompts Works with Claude + ChatGPT
Something went wrong. Please try again.

No spam. Unsubscribe anytime.

The unlazy repo + the two prompts that make it work Jump to the fix → Join the Membership
Wright Mode — Free Resource

The fix for AI saying it's done when it isn't

Claude says the work's finished, you check, three steps are missing. The unlazy skill fixes "partial compliance" — here's the repo and the prompts.

🆓 Free GitHub skill ✅ Checklist + definition of done ⚡ Parallel agents prompt included

What's inside

1The repo 2The prompts 3Why it works

The repo (start here)

The problem has a name: partial compliance. The AI works for 30 minutes, tells you it's done, and it's quietly skipped steps. A developer built a skill that fixes it.

unlazy

github.com/Leonxlnx/unlazy — the anti-laziness skill. Every project gets broken into smaller tasks, each with its own checklist and one line saying what finished actually looks like.

The two prompts

Prompt 1 — install it. Paste this into Claude Code:

Install the unlazy skill from https://github.com/Leonxlnx/unlazy and set it up. From now on, when I give you a project, break it into smaller tasks — each with its own checklist and one line defining what finished actually looks like. Don't report anything as done until every box on its checklist is genuinely ticked.

Prompt 2 — keep it fast. This is the important one from the video — without it, all that checking makes big projects take hours:

You can run multiple agents at the same time to work through these checklists. Split the task list between parallel agents wherever tasks don't depend on each other, then bring the results together and verify every checklist before telling me the project is done.

Why it works

"Done" is fuzzy, and AI exploits fuzzy. A checklist with a written definition of finished — "done = deployed and tested", not "done = I wrote some code" — removes the wiggle room. The skill just makes that discipline automatic on every task, so you stop being the quality inspector.

Ready to go deeper?

Take your skills to the next level with these resources.

Community + Training

Wright Mode Membership

Join a community of women entrepreneurs implementing AI and automation in their businesses. Live calls, templates, and ongoing support.

Learn more →
Workshop Replay

Claude Masterclass

Learn how to use Claude like a pro. From prompting fundamentals to building real workflows that save hours every week.

Watch now →
Next Level

Claude Code Masterclass

Go beyond the chat interface. Build automations, process data, and create tools with Claude Code — no developer experience needed.

Explore →