* feat(skills): add hyperframes-cli skill for CLI workflow guidance Adds a new skill that teaches AI agents how to use the HyperFrames CLI (init, lint, dev, render, doctor). Previously, agents had no way to discover the CLI — the compose-video skill only covered HTML authoring. This led to agents searching for binaries, finding the monorepo, and running bun run studio manually instead of using npx hyperframes dev. Also registers the skill in init.ts so new projects get it bundled alongside hyperframes-compose and hyperframes-captions. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * refactor(cli): rename dev command to preview The command starts a preview server — "preview" describes what users are doing more accurately than "dev". Updates the command name, file name, all CLI references, docs, skills, and template CLAUDE.md. 22 files updated across CLI source, docs, skills, and templates. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(skills): replace stale dev reference with preview in CLI skill Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(docs): catch remaining dev references missed in rename - testing-local-changes.mdx: two inline command examples - troubleshooting.mdx: anchor link #dev → #preview, "dev server" → "preview server" - cli.mdx: "dev server" → "preview server" Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
4.9 KiB
Documentation Guidelines
Standards for writing and maintaining Hyperframes documentation. Based on patterns from Remotion, Stripe, Tailwind CSS, and Astro.
Core Principles
-
One-sentence intro rule — Every page opens with a single sentence telling the reader what this page helps them do or understand. No preamble, no history.
-
Outcome before implementation — Show what the code produces (rendered result, terminal output, file structure) before showing the code itself.
-
Show, don't tell — Use concrete examples with realistic values. Never use
foo/bar/baz. Prefer a working HTML snippet over a description of what to write. -
Two content modes — Guides build narratives with progressive complexity. References enable scanning with standardized structure. Never mix them.
-
No dead ends — Every page links forward (next steps), backward (prerequisites), and sideways (related concepts). Readers should never reach a page with nowhere to go.
Page Structure
Guides (concepts/, guides/)
Title
├── One-sentence purpose statement
├── What this looks like (output, demo, or visual)
├── Minimal working example
├── Deeper explanation with progressive complexity
├── Common patterns / best practices
├── Warnings and pitfalls (sparingly)
└── Next steps (cards or links to related pages)
Reference pages (reference/)
Title
├── One-sentence definition
├── Complete attribute/API table
├── Detailed section per item (type, default, description, example)
├── Rules and constraints
└── Related pages
Package pages (packages/)
Title
├── One-line description + install command
├── When to use this package (and when NOT to)
├── Key features list
├── Minimal usage example with expected output
├── Configuration reference
└── Related packages
Writing Style
- Second person, active voice, imperative mood: "Use X to do Y." Not "The developer should consider using X."
- Present tense: "The runtime manages media playback." Not "The runtime will manage..."
- Be direct: "This breaks rendering." Not "This may potentially cause issues with the rendering pipeline."
- Prerequisites at point of need: State requirements where they matter, not in a wall at the top.
- Conversational but precise: Friendly tone, exact technical details.
Code Examples
Always annotate code blocks
```html index.html
<div data-composition-id="root" ...>
```
The filename after the language tag tells readers where the code goes.
Use numbered comments for multi-step code
// 1. Create a paused timeline
const tl = gsap.timeline({ paused: true });
// 2. Add animations
tl.from("#title", { opacity: 0, y: -50, duration: 1 }, 0);
// 3. Register the timeline
window.__timelines["my-video"] = tl;
Show expected output
After CLI commands, show what the user should see:
npx hyperframes preview
# ✓ Server running at http://localhost:3000
# ✓ Watching for changes...
Use CodeGroup for multi-platform commands
<CodeGroup>
```bash macOS
brew install ffmpeg
```
```bash Ubuntu
sudo apt install ffmpeg
```
</CodeGroup>
Mintlify Components — When to Use
| Component | Use When |
|---|---|
<Steps> |
Sequential setup or tutorial instructions |
<CodeGroup> |
Same action across platforms/languages |
<Tabs> |
Alternative approaches with equal weight |
<Card> / <Columns> |
Navigation to related pages, next steps |
<Accordion> |
FAQ or optional detail that would bloat the page |
<Note> |
Non-obvious behavior the reader should know |
<Warning> |
Something that will break if ignored |
<Tip> |
Helpful shortcut or best practice |
<Info> |
Context that aids understanding |
<Tree> |
File/directory structure |
<Frame> |
Screenshots or diagrams with captions |
Callout budget: max 2-3 per page
More than 3 callouts creates alert fatigue and readers skip them all. Reserve <Warning> for things that genuinely break. Use inline prose for tips.
Cross-Linking
- Link at the point of curiosity: When you mention a concept that has its own page, link it immediately. Don't hoard links.
- "See also" at page bottom: Only for genuinely related content that doesn't fit inline.
- Next steps cards: End guide pages with
<Card>links to logical next pages.
File Conventions
- All doc pages are
.mdx(not.md) - Use kebab-case for filenames:
frame-adapters.mdx, notframeAdapters.mdx - Frontmatter requires
titleanddescription - Description should be under 160 characters (used for SEO/social)
Maintenance
- Docs live in the repo at
/docsand deploy automatically on merge tomain - PRs that change user-facing behavior should update relevant doc pages
- Run
mint validateandmint broken-linksbefore pushing doc changes