Land
Stay present with AI. A Claude Code plugin that keeps you understanding what you sign to own.
npx -y skills add saadshahd/moo.md --skill landAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
What its author says it does
Copied from the file, not written here
Use when the user wants knowledge delivered to an audience — slides, a deck, a diagram, an explainer for others, teaching or talk material — or shares existing content that isn't getting through.
SKILL.md
2.9 KB, as published. Nobody here has run it
Deliver information the way minds absorb it. Apply silently; explain the reasoning in plain language only when the user asks why.
Before designing
Ask only what's missing, then confirm the brief — audience and the one thing — before generating:
- Audience — sets language, analogies, assumed prior knowledge.
- The one thing — if they keep a single memory, action, or feeling, what is it? Everything serves it.
Open with a hook that creates a curiosity gap — a question, a surprising number, a misconception flipped — never a topic title.
Principles
- One cognitive unit at a time. One idea per slide, section, or screen. Split rather than crowd.
- Visuals carry the information. Acceptance test: delete every word — the meaning must survive. Text is a label, not the explanation. Draw the concept itself: arrow length = force, area = fraction, scene density = era.
- Annotations live inside the graphic, attached to the structures they name — never a list beside the picture.
- Change is shown as motion or sequence — process, relation, and comparison are animated or stepped, paced by the audience, replayable. Nothing autoplays.
Knowledge type → representation
| The content is... | Show it as |
|---|---|
| Concept ("what is X") | An everyday analogy drawn as a graphic, correspondences mapped across (API = restaurant ordering) |
| Procedure ("how to") | Steps revealed one at a time, prior steps stay visible; branches as a flow diagram |
| Relation ("X and Y") | Nodes with drawn connections, Venn, or 2×2 matrix — line weight = strength |
| Story / history | A growing timeline, or before/after scenes where the scenery itself is the data |
| Numbers | Area, count, or proportion anchored to a familiar referent — never a bare table |
| Misconception | Before/after flip: show the wrong belief first, then the correction |
Defaults — the user's call, not yours
When the deliverable is a deck and the user hasn't specified: one self-contained HTML file; fixed canvas 1920×1080 scaled to the viewport with transform: scale() (not vw/vh); click/space/arrow advances step reveals, then slides; one coherent theme (~5 colors, 2 fonts) matched to the audience; text legible on a phone; nothing overflows the canvas.
Never
- Bullet-point lists as the design — reach first for a layout, diagram, or step reveal, and keep the list only where a diagram would merely decorate it.
- The narration script pasted onto the artifact — spoken words and shown visuals carry different information.
- Dark-gradient-plus-white-text "AI deck" aesthetic.
- Animation on a timer — a
setIntervalkeeps firing after the slide changes, so whatever it drives outlives the slide that started it.