Skip to content
Pi Extensions & Skills

Language & platform skills

Skills shine for stack-specific knowledge — Apple’s latest UI frameworks, a team’s coding rules, framework conventions. Because skills load on demand, you can carry a lot of reference material without bloating every prompt.

Alternatives

Section titled “Alternatives”

1. macos-dev-code (skill)

Section titled “1. macos-dev-code (skill)”

A skill for native macOS development with SwiftUI, AppKit, and Liquid Glass. It bundles curated Apple Developer documentation (snapshotted), so Pi follows current Apple conventions instead of guessing.

Terminal window
pi install npm:macos-dev-code

Benefits

  • Up-to-date, curated platform docs baked in.
  • Keeps the model on idiomatic, current APIs.
  • Loaded on demand — no constant prompt cost.

Drawbacks

  • Scoped to Apple/macOS development.
  • Bundled docs are a snapshot; they age between updates.

2. @nitra/cursor (rules loader)

Section titled “2. @nitra/cursor (rules loader)”

A CLI/package that downloads Cursor rules (prefix n-) into your local repo, letting you reuse the large ecosystem of Cursor .mdc rule files as Pi guidance.

Terminal window
pi install npm:@nitra/cursor

Benefits

  • Taps the existing Cursor rules ecosystem.
  • Brings team/framework conventions into the repo.

Drawbacks

  • Cursor rules aren’t native Pi skills — expect some impedance mismatch.
  • You manage which rules land in the repo.

3. Write your own project skill (no package)

Section titled “3. Write your own project skill (no package)”

The most precise option: a SKILL.md in .pi/skills/ capturing your project’s conventions, commands, and gotchas.

.pi/skills/our-stack/SKILL.md
# Our Stack
Use when working in this repo.


## Conventions
- Use pnpm, not npm.
- All API handlers live in `src/api/` and return typed results.

Benefits

  • Perfectly tailored; lives with your code in version control.
  • Reuses Claude Code / Codex skills too (Pi can read their directories).

Drawbacks

  • You write and maintain it.
  • No external curation.

Which should I pick?

Section titled “Which should I pick?”
  • Doing macOS/Apple work: macos-dev-code.
  • Want to reuse Cursor rules: @nitra/cursor.
  • Encoding your own repo’s conventions: write a project skill in .pi/skills/.