Pi Extension Development logo

Pi Extension Development

Community
HazAT
pi-extension-development

Defines the local standard for standalone Pi extension repositories. Use when manually invoked to create, audit, test, publish, or release an extension with Git-backed installation, local loading, and Vite+ linting and formatting.

Overview

PublisherHazAT
Repositorypi-config
Skill namepi-extension-development
Stars
450
Forks
44
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by HazAT on GitHub. Read the source before you install it.

Installation

Install the Pi Extension Development AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/HazAT/pi-config.git /tmp/pi-config
mkdir -p .claude/skills
cp -r /tmp/pi-config/skills/pi-extension-development .claude/skills/pi-extension-development
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Pi Extension Development in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Pi Extension Development on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Pi Extension Development is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Pi Extension Development

Build standalone Pi extensions as Git-backed Pi packages with consistent local development, verification, and release workflows. Never publish these repositories to npm.

Step 1: Establish the task

Interpret the arguments appended to the skill invocation as an operation and optional path:

  • create <path-or-name> — scaffold a standalone extension repository.
  • audit <path> — compare an existing repository against this standard.
  • release <version> — apply the standardized release workflow to the current repository.
  • No operation — ask what repository and outcome the user wants.

Before creating files, inspect the target path and read its AGENTS.md, CLAUDE.md, and applicable rule files. Never overwrite an existing repository or unrelated work.

Step 2: Read only the needed reference

OperationRead
Create or restructure a repositoryreferences/repository-structure.md
Implement or review extension codereferences/extension-guidelines.md
Release or add a release skillreferences/release-workflow.md

For Pi APIs, read the installed Pi documentation completely before implementation:

  • Extensions: /opt/homebrew/lib/node_modules/@earendil-works/pi-coding-agent/docs/extensions.md
  • Packages: /opt/homebrew/lib/node_modules/@earendil-works/pi-coding-agent/docs/packages.md
  • Skills: /opt/homebrew/lib/node_modules/@earendil-works/pi-coding-agent/docs/skills.md
  • TUI, when needed: /opt/homebrew/lib/node_modules/@earendil-works/pi-coding-agent/docs/tui.md

Follow relevant links and examples referenced by those documents.

Step 3: Verify the development tools

Use the globally installed Vite+ CLI:

bash
command -v pi
command -v vp
vp --version

If vp is absent, report that and stop before adding a substitute formatter or linter. Do not install standalone Oxlint, Oxfmt, ESLint, Prettier, or Vitest when Vite+ provides the needed command.

Use Vite+ for package operations and checks:

bash
vp install
vp add -D vite-plus
vp check
vp check --fix
vp fmt --check
vp lint

The global vp executable drives the workflow. Keep vite-plus in devDependencies because vite.config.ts imports defineConfig; Git package production installs omit it.

Step 4: Use the canonical package model

Require:

  • A standalone Git repository under ~/Projects/ unless the user specifies another location.
  • A package.json with private: true, keywords: ["pi-package"], peer dependencies for Pi-provided imports, and an explicit pi manifest.
  • TypeScript source loaded directly by Pi; do not add a build or bundled dist/ directory unless runtime requirements make one necessary.
  • Installation through pi install git:github.com/OWNER/REPO, never npm publication.
  • A checked-in .pi/settings.json that disables an installed copy while loading checkout-local resources.
  • Vite+ configuration, editor settings, and a checked-in vp staged pre-commit hook.
  • A manual-only repository release skill under .pi/skills/release/.

Step 5: Implement the minimum extension

Prefer one focused extension entry point and split files only when the implementation has real independent concerns. Reuse Pi APIs and Node built-ins before adding dependencies.

For custom tools:

  • Use strict TypeBox schemas and StringEnum for string enums.
  • Set executionMode: "sequential" for stateful, desktop-control, or otherwise conflicting operations.
  • Honor AbortSignal and set bounded timeouts for subprocesses or network work.
  • Truncate output to Pi limits and preserve full output in a temporary file when truncated.
  • Throw errors to produce failed tool results.
  • Return native image blocks for generated screenshots or previews.
  • Start long-lived resources only after session_start or on demand, and close them during session_shutdown.

Do not request permissions, perform destructive setup, or mutate unrelated user configuration without explicit approval.

Step 6: Configure local loading

Verify both paths:

bash
# Explicit checkout loading
pi -e ./pi-extension/EXTENSION/index.ts --skill ./skills/SKILL/SKILL.md

# Git installation
pi install git:github.com/OWNER/REPO

When developing inside the checkout, .pi/settings.json must prevent duplicate global and local resources.

Step 7: Verify before publishing

Run, at minimum:

bash
vp check
pi --no-extensions -e "$PWD" --no-skills --list-models '__load_check__'
git diff --check
git status --short

Also execute the registered tool through Pi's extension loader or a fresh Pi session with a safe real scenario. Verify failure propagation, cancellation or timeout behavior, truncation when applicable, and cleanup of temporary artifacts.

Step 8: Publish and install

Before any commit, read and follow the global commit skill. Create a focused conventional commit. Create or push a public GitHub repository only when the user authorizes publication.

After pushing, install from Git and verify the installed checkout rather than relying only on source-tree tests:

bash
pi install git:github.com/OWNER/REPO
pi --no-context-files --list-models '__installed_load_check__'

Report repository URL, commit SHA, install command, verification evidence, and any remaining manual permission step.

Exit criteria

Finish only when:

  • vp check passes.
  • Pi loads the local package and the Git-installed package.
  • The repository and installed checkout are clean.
  • No temporary test files remain.
  • Documentation contains install, local development, permissions or prerequisites, verification, and release instructions.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Pi Extension Development AI skill do?

Defines the local standard for standalone Pi extension repositories. Use when manually invoked to create, audit, test, publish, or release an extension with Git-backed installation, local loading, and Vite+ linting and formatting.

Why use Pi Extension Development on TypingMind?

Because you install it once and use it with any model. Pi Extension Development is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Pi Extension Development in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/HazAT/pi-config/tree/main/skills/pi-extension-development. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Pi Extension Development?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Pi Extension Development?

As many as you like. As long as a model supports skills, you can use Pi Extension Development with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Pi Extension Development AI skill free?

Yes. It is published on GitHub by HazAT under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇