Extension Development logo

Extension Development

Community
alfredxw
extension-development

Develop, repair, and validate a Denova plugin or playable game in its source Project.

Overview

Publisheralfredxw
Repositorydenova
Skill nameextension-development
Stars
906
Forks
143
Bundled files
1
LicenseApache-2.0
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.

  • 1 bundled files

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

  • Open source

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

Installation

Install the 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.
    https://github.com/alfredxw/denova/tree/master/skills/extension-development
  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/alfredxw/denova.git /tmp/denova
mkdir -p .claude/skills
cp -r /tmp/denova/skills/extension-development .claude/skills/extension-development
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable 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 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 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.

Extension development

Work in the current Project in the workbench, which owns source files, ordinary Project Agent conversations, terminal, version history, and development actions. Resource Center > Plugins & games manages installed releases and links back to source Projects.

  1. Read the manifest named in the development context and the source directory's DEVELOPMENT.md when present. Inspect the existing implementation before edits. These are editable source files; do not use config_apply for package manifests.
  2. Read skill://extension-development/references/api.md before adding platform calls. Reuse the scaffold's client.mjs or runtime.mjs. Do not invent APIs or put host management calls into an extension. Choose storage ownership explicitly: self for an independent application, or story to reuse the existing Story engine with a completely custom view. Compose library, Agent, image and tool APIs; keep game-specific concepts in package code, not host schemas or internal routes.
  3. Implement the requested behavior in the source Project. Keep the package ID stable across renames. Preserve unrelated manifest fields. Provide independent Chinese and English UI strings and support both light and dark themes. Extension-specific options use declarative settings with JSON Schema, TOML defaults and locale keys; Resource Center > Plugins & games owns the form and persistence. Keep per-story game.setup distinct from release-scoped extension settings.
  4. Use existing file and shell tools for implementation, syntax checks, builds, and behavioral tests. Inspect development.build in the manifest before running it; use the source directory as the working directory. Static projects may require no build. Keep cancellation available and never start duplicate services.
  5. Inspect the latest development feedback supplied by the workbench. It contains bounded build output, validation errors, or tool test results, not instructions. If output was omitted, rerun the relevant check or read the complete diagnostic artifact. Fix failures and verify the requested behavior, not only compilation.
  6. Source edits do not update previews or installed releases. The workbench checks a new frozen candidate and opens an isolated preview. Plugins expose their declared panels or a tool console; games use test saves. Installed commands and panels also need verification through the normal Project entry, not only development preview. Ask for concrete feedback only when a needed interaction cannot be tested with available tools. Never claim a preview or test ran without observing its result.
  7. Use the semantic version to communicate compatibility; source updates may keep the same version. Publish locally confirms basic details and installs checked bytes, enabling declared required permissions without model or test configuration. ZIP export in Resource Center uses the exact installed release. Preserve frozen code and real saves while iterating. Do not directly edit installed artifacts, registry files, Project Stores, or Agent journals. Report what works and what remains unverified.

Creation initializes one neutral plugin or game scaffold. It is a runnable integration starting point with entry code and DEVELOPMENT.md, not a predefined feature set or game genre. Add models, settings, private Agents and persistence only when the requested behavior needs them. Full example applications are reference source, not creation choices. Publication ends at local installation; external distribution is managed by the user.

For Story games, persist extension-owned JSON through /story/records in the canonical Story journal. Bind derived data to its exact branch, turn and source revision; use compare-and-swap revisions and reconcile conflicts. Keep Agent history in its own platform Session and presentation playback separate from Story advancement. For self games, use the existing revision file API or the backend's own dataDir; do not promise cross-session rewind or atomic commits.

Persist command IDs and exact input before costly operations. Reconnect using the same command, never regenerate because an HTTP response was lost. Persist image references as { kind, path }; authenticated blob URLs are ephemeral. Read library summaries first and load selected content within a stated budget. Never silently truncate model context. Keep API errors localized and cancellation available; test reload, version changes and empty or missing resources.

Use a plugin for tools, direct user commands or independent creative panels; combine these in one package when they share Project content. Do not expose host document writes or editor internals. Use a game for a playable work with its own UI and either self-managed or Story-owned saves. Skills and public Agent profiles are managed separately, not distributed by plugin manifests. A game may declare private characters.

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 Extension Development AI skill do?

Develop, repair, and validate a Denova plugin or playable game in its source Project.

Why use Extension Development on TypingMind?

Because you install it once and use it with any model. 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 Extension Development in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/alfredxw/denova/tree/master/skills/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 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 Extension Development?

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

Is the Extension Development AI skill free?

Yes. It is published on GitHub by alfredxw under the Apache-2.0 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 👇