Learn

An Agent Skill is a folder with a SKILL.md file that your coding agent loads on demand. The format, where it comes from, which agents read it, and what PalateKit adds around it.

Pick a PalateKit resource, install it in your coding agent in a minute, then check that the agent uses it.

Install prompt, npx command or ZIP: when to pick each, where files go for each resource type, project or user level, and signing the CLI in for paid resources.

Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI or OpenCode: what each one reads (skills, AGENTS.md, rules) and where it looks for files.

How to phrase a request, let the agent pick the skill or call it by name, combine several skills and adapt one to your project.

How an agent decides to load a skill, why a skill is missed or fires too often, and how to write a description that triggers at the right moment.

List your installed skills, update them, disable or remove them, and avoid duplicates between your user folder and the project folder.

What a skill can make your agent do, how to read one in two minutes, which permissions to keep, and what PalateKit checks in its catalog.

What a DESIGN.md file contains, where to put it, how to make your agent follow it and how to adapt it to your brand.

What a catalog motion template contains, how to add it to your project or have your agent adapt it, render video with HyperFrames and respect prefers-reduced-motion.

Paste an image or video prompt into your generator, replace the variables, iterate on the result, and know who owns the images it produces.

Rules, subagents and hooks in Claude Code, Cursor and Codex: what each one does, which folder it goes in, how it differs from a skill, and what to check before enabling a hook.

The seven most common problems after installing a resource (skill missing, never triggered, file not found, npx command failing, badly extracted ZIP, conflict, paid resource refused), each with its likely cause and a fix.

What a skill, a DESIGN.md or a rule adds over a prompt pasted into the chat, and when a prompt is enough.

What a skill does, what an MCP server does, how much context each costs, their risks, and how to combine them in one project.

CLAUDE.md, AGENTS.md, Cursor rules, Copilot instructions or a skill: what is always loaded, what loads on demand, and how to split your guidance.

The open Agent Skills standard, official repositories, catalogs, Claude Code plugins and neighboring formats (DESIGN.md, AGENTS.md, MCP): who does what, and where PalateKit fits.

React or HTML/CSS, license, rendering, use cases: how to choose between Remotion and HyperFrames when a coding agent writes the video.

Frontmatter fields, naming rules, folder structure, size and progressive loading of an Agent Skill, with a minimal example.

The skill folders (user and project) and instruction files of Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI and OpenCode, based on their official documentation.

What Claude produces (text, code, SVG, HTML/CSS), what it does not (photos, raster illustrations), and how to use an image generator with the catalog's resources.

What a design skill adds to the interfaces your agent builds (typography, palette, layout), how to choose one, pair it with a DESIGN.md and ask for it.

HTML slides, PDF or PowerPoint export, libraries and pptx skills: how to produce a slide deck with your agent, and when a dedicated presentation tool is the better choice.