skills.rest
popmechanic
popmechanicCOMMUNITY
Explore all skills by this author →
GitHub

loom

Orchestrates Claude Code runtime to build interactive applications via REST, SSE, and WebSocket.

3|Updated Mar 1, 2026
Search Tags:#oauth#runtime#streaming#nodejs#claude-code#websocket#loom
One-click install
npx skills add https://github.com/popmechanic/loom --skill loom
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: loom
Source: https://github.com/popmechanic/loom/tree/main/skills/loom
Command: npx skills add https://github.com/popmechanic/loom --skill loom

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Loom enables developers to build applications where Claude Code runs as the runtime. It provides a server that spawns claude -p processes or uses the Agent SDK, a front-end interface, and streaming capabilities to render structured outputs in real time.

Core Features & Use Cases

  • Architecture that connects a frontend UI to a Node.js server and a Claude runtime, enabling file access, command execution, and streaming results.
  • Supported patterns include REST + SSE for one-shot tasks, WebSocket for multi-turn sessions, and OAuth-based multi-user authentication.
  • Use cases span interactive dashboards, code analysis pipelines, and automated tool orchestration with structured outputs.

Quick Start

Start a Loom app by configuring a frontend to communicate with a Node.js server that spawns claude -p processes and streams structured results.

Frequently Asked Questions about loom

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run Claude Code as a backend runtime for my Node.js application?▼

You can run Claude Code as a backend runtime by using a Node.js server to spawn `claude -p` processes or the Agent SDK, then streaming structured results to your frontend interface for real-time rendering.

What is the best way to stream Claude Code responses to a frontend UI?▼

Streaming Claude Code responses to a frontend is best handled via REST with SSE for one-shot tasks or WebSocket connections for multi-turn interactive sessions, coordinating tools and rendering structured outputs.

Do I need OAuth to use Claude Code in a multi-user web application?▼

Yes, you need OAuth setup for Anthropic credentials to authenticate users when building multi-user applications that spawn the Claude Code runtime and stream results across your server.

Can I use the Agent SDK to orchestrate interactive dashboards with Claude?▼

Yes, you can use the Agent SDK to orchestrate interactive dashboards by spawning Claude Code processes on a Node.js server and streaming structured outputs to your frontend interface.

When should I use WebSocket instead of SSE for Claude Code streaming?▼

You should use WebSocket instead of SSE when building multi-turn interactive sessions with Claude Code, whereas REST plus SSE is sufficient for one-shot task execution and streaming results.

What limitations exist when building apps that spawn Claude Code processes?▼

Limitations include requiring a Node.js or TypeScript server environment, OAuth setup for Anthropic credentials, and a dedicated frontend to render the structured outputs streamed from the spawned Claude processes.

Related Skills

Explore More Software Engineering Skills→
hhopkins95hhopkins95
2

Backend Setup

Configure a Node.js agent-runtime backend with REST and WebSocket servers.

Community
Advanced
BewinxedBewinxed
106

river-ts-streaming

Unify typed SSE and WebSocket streaming with river.ts.

Community
Advanced
LangbaseIncLangbaseInc
103

building-ai-agent-on-cloudflare

Build scalable, stateful AI agents on Cloudflare with Durable Objects and WebSockets.

Official
Advanced
alxxpersonalalxxpersonal
12

claude-headless

Control Claude Code headlessly via NDJSON streams for custom interfaces.

Community
Advanced