Your agent can build a site, a deck, a video, an image. It just has no taste.
Skillry is a growing library of Agent Skills that do. Each Skill has a specific look, and you can see exactly what it makes before you install it — you are not buying a description.
Install from the web or your CLI, then invoke a Skill by name inside your agent and get a finished piece of work: a landing page that looks art-directed, a deck you would actually present, an image that stops the scroll, a short film that moves.
Works with Claude Code, Codex, Cursor and a dozen other agents.
The taste is ours. The work is your agent's.
Features
See exactly what a Skill makes before you install it
One-command install from the web or your CLI — no API key
Four kinds of work: web pages, slide decks, images and video
Works with Claude Code, Codex, Cursor and a dozen other agents
New Skills added every week
Use Cases
Ship a landing page that looks art-directed instead of templated
Turn a rough brief into a deck you would actually present
Make covers and OG images that share one visual system
Produce a short launch film without opening an editor
Give a coding agent a house style it keeps across every output
Comments
The output previews make the choice easier. If someone uses one Skill for a landing page and another for an OG image or launch video, can they share colors and typography across those outputs, or are the styles independent today? Cross-format consistency is where agent-built launch assets often drift.
I build with Claude Code every day. It writes good code. Then it ships a landing page with a purple gradient, a deck with four fonts, and an OG image that looks like every other OG image on the internet. The code was never the problem. The taste was. So we built Skills with an actual point of view. Each one has a specific look, and you can see exactly what it makes before you install it — you're not buying a description. Browse the output, then install to your own agent with one command. Claude Code, Codex, Cursor and a dozen others. The taste is ours. The work is your agent's. Four kinds of work so far: web pages, slide decks, video, images. The library grows every week. Honestly, what it isn't: it's not a marketplace, there's no community upload, and video is the thinnest category right now. It's opinionated — if our taste isn't yours, you'll feel that in the first five minutes, and I'd rather you find out fast. Tell me what your agent keeps getting wrong. That's the request queue we actually build from.
Letting people see exactly what a Skill produces before installing it is the right call; "trust the description" is how most prompt libraries lose people. The line about the taste being yours & the work being the agent's sums up the gap with generic AI output well. Are the Skills versioned, so a design you like doesn't shift under you after an update?
Premium Products
Sponsors
Buy
Makers
Makers
Comments
The output previews make the choice easier. If someone uses one Skill for a landing page and another for an OG image or launch video, can they share colors and typography across those outputs, or are the styles independent today? Cross-format consistency is where agent-built launch assets often drift.
I build with Claude Code every day. It writes good code. Then it ships a landing page with a purple gradient, a deck with four fonts, and an OG image that looks like every other OG image on the internet. The code was never the problem. The taste was. So we built Skills with an actual point of view. Each one has a specific look, and you can see exactly what it makes before you install it — you're not buying a description. Browse the output, then install to your own agent with one command. Claude Code, Codex, Cursor and a dozen others. The taste is ours. The work is your agent's. Four kinds of work so far: web pages, slide decks, video, images. The library grows every week. Honestly, what it isn't: it's not a marketplace, there's no community upload, and video is the thinnest category right now. It's opinionated — if our taste isn't yours, you'll feel that in the first five minutes, and I'd rather you find out fast. Tell me what your agent keeps getting wrong. That's the request queue we actually build from.
Letting people see exactly what a Skill produces before installing it is the right call; "trust the description" is how most prompt libraries lose people. The line about the taste being yours & the work being the agent's sums up the gap with generic AI output well. Are the Skills versioned, so a design you like doesn't shift under you after an update?
Premium Products
New to Fazier?
Find your next favorite product or submit your own. Made by @FalakDigital.
Copyright ©2026. All Rights Reserved