
You give Claude a folder of your footage and music plus your style rules, then save the style it learns. After that you drop in a raw one-take video and Claude cleans the sound, cuts the mistakes and edits it your way.
Claude CodeSkills by use case, each with what it does and what it needs.
321 skills, sorted by top

You give Claude a folder of your footage and music plus your style rules, then save the style it learns. After that you drop in a raw one-take video and Claude cleans the sound, cuts the mistakes and edits it your way.
Claude Code
Builds polished, easy-to-use web page elements for you — buttons, forms, pop-ups, dark mode and more — using ready-made design pieces, so your app looks professional without you styling everything by hand.
Claude Code
Walks you through finding local businesses worth teaming up with by spotting shared customers in reviews, then writes the partnership pitch and a ready-to-send first email.
Apify
Google Workspace
Notion
Google Places APIHeads up: Orbit flagged something to check before you run it. See the full page.

After Claude gives you advice, a plan, or numbers you might act on, it tacks on a couple of quick questions to help you double-check the facts and spot anything missing — just once per chat, and it skips simple lookups or code requests.
Claude Code
Takes a video and automatically burns in those bouncing, word-by-word highlighted captions you see on TikTok and Shorts, and can even copy the caption style from a screenshot you give it.
Claude Code
Shows you how to hand Claude a whole piece of work with full context so it plans and finishes the job on its own, instead of you managing it step by step.
Claude Code
Acts like a startup finance chief. You share your cash, revenue and customer costs, and it works out how long your money will last, whether you're spending too fast, and how fundraising would shrink your ownership.
Claude Code
Acts as a front desk for the gstack set of AI workflows, picking which one fits your task. You get sent to the right workflow instead of guessing which to use.
Claude CodeHeads up: Orbit flagged something to check before you run it. See the full page.

Lets one main agent run a team of coding agents. The team members report when their work is ready, and the main agent checks it, merges it in batches, and writes a summary of progress.
Claude CodeHeads up: Orbit flagged something to check before you run it. See the full page.

When a test fails, code crashes, or output looks wrong, this finds the real cause of the problem and then fixes it. You get a short report with the cause, the fix, and proof it works.
Claude CodeHeads up: Orbit flagged something to check before you run it. See the full page.

Asks you the right questions before anything gets built, so the AI understands what you really want and why. You end up with a short, approved design doc and a plan to build from.
Claude Code
Has the AI carry out a written plan step by step in one session, testing each step as it goes. You end up with the finished work, a progress record, and one final review.
Claude Code