Created two skills a. draft-newsletter b. save-the-newsletter bundled together to a plugin. output: AI in QA Newsletter.

Lesson 6.2 Download newsletter-skill/ Run it once. Pick any topic that matters to you — your favorite breakfast, the last book you finished, what you wish your in-laws understood about your job, why you stopped using a tool you used to love. Save the output as your first draft. Do I change line 4 in SKILL.md? (allowed-tools) to My Favorite ____ Food? I did that, now what is the next step? I don't even see /draft-newsletter as a claude command where I think it should be. I see /hello-name and /deep-research and /add-dir. I also see in VSCode > Source Control > 10000 pending changes which seems like a result of what I did last week to fix the yml file. If I commit these changes, how big a number of token will that require?

Here is the part that still surprises people. With one folder of skills and a Claude Code Pro subscription, you can build almost anything you can describe. A real tool. A slice of your weekly work, done for you. Even a small business. No frameworks, no computer science degree. Just a folder you open and run. But none of that lands until your environment is set up and Claude Code is actually running on your own machine. So before we go further together, I want to know exactly where everyone is right now. Vote below. And whatever you pick, you are not doing this alone. Drop a comment telling us where you are, or where you got stuck, and we will help you get running. That is the whole point of this room. No rush and no judgment. We start from wherever you are. Pragati

*Context* I manage multiple product teams at a large enterprise SaaS company. Every week we do architecture reviews - reading docs, analyzing code patterns, checking for anti-patterns and optimization opportunities. It was eating multiple hours of my time, and in some cases i feel a lot unprepared for meetings. *What I built* A Claude Code skill that handles the entire review workflow: - Reads architecture documentation automatically - Analyzes code patterns across multiple repos - Identifies optimization opportunities - Checks for common anti-patterns - Generates summary reports with recommendations *Results* - ⏱️ 4 hours → 30 minutes (>85% time saved) - 📊 More consistent reviews (no human fatigue) - 🎯 Team can focus on decisions, not analysis (It really helps to read the summary) - 🔄 Runs weekly automatically *Key lesson* Don't just automate tasks - automate entire workflows. This skill handles the full review process from start to finish, not just one step. That's where the real leverage is. *For the technical folks* The skill uses: - Document parsing for architecture docs - Code analysis across repos - Pattern matching for anti-patterns - Structured report generation - Integration with our internal docs system Happy to share the approach if anyone's tackling similar workflows at enterprise scale.

This is the most important category in the community. Every build, every revenue screenshot, every "holy shit it worked" moment goes here. Post your wins. No win is too small. "I built my first SKILL.md" is a win. "Claude Code automated my invoicing" is a win. "I made $200 this week with a skill I built" is a win. House rule: every new win gets pinned for 48 hours. Get featured on YouTube if it's exceptional. Screenshots + before/after + what skill you used = instant feature.