mikker/holepunch-skills · Archived

holepunch-app-composition

Compose a complete Holepunch application by assigning responsibilities across the host shell, Pear runtime, backend worker, shared domain layer, RPC boundary, and UI state.

First seen Jun 21, 2026

Installation

$ npx skills add mikker/holepunch-skills --skill holepunch-app-composition

Summary

  • Compose a complete Holepunch application by assigning responsibilities across the host shell, Pear runtime, backend worker, shared domain layer, RPC boundary, and UI state.
  • Use for tasks that plan or refactor app architecture, decide where features should live, connect Hyperstack primitives to Pear hosts, or turn a product idea into a practical implementation blueprint without coupling it to one specific app.

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from mikker/holepunch-skills.

npx skills add mikker/holepunch-skills

Browse all from mikker/holepunch-skills

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 1
Default branch main
Open issues 0
Status Archived

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,012 B
  • docs SUMMARY.md 445 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 1 installs

SKILL.md

Holepunch App Composition

Overview

Use this when the problem is bigger than one library. The goal is to produce a clean architecture where the runtime boots the worker, the worker owns replication, the domain layer owns rules, and the UI consumes snapshots.

Preferred Shape

  • Host shell: windowing, preload bridge, runtime lifecycle, update events.
  • Worker or service: long-lived backend, managers, watchers, snapshots, file ingestion.
  • Domain layer: contexts, schemas, permissions, replicated data model.
  • UI layer: state store, optimistic UX where safe, rendering only.

Architectural Rules

  • Keep app-specific business rules out of the host shell.
  • Do not let the renderer talk to Corestore or Hyperswarm directly if a worker can own that state.
  • Build one RPC surface per worker and keep it task-oriented, not storage-oriented.
  • Let the worker stream snapshots or events to the UI instead of exposing many low-level queries.

Compose in This Order

  1. Define the replicated domain and local metadata split.
  2. Choose the host and runtime wiring.
  3. Define the worker entrypoint and ownership of managers and watchers.
  4. Define the RPC surface the UI actually needs.
  5. Shape the UI around snapshots, subscriptions, and task-level mutations.
  6. Add test seams at the worker and domain boundaries first.

Load References When Needed

  • Read references/blueprint.md for the recommended layer boundaries and build order.
  • Read references/examples.md for example architectures that have already proven workable.