setup-sandcastle

Orchestrate autonomous agents across sandboxed environments for parallel GitHub issue tasks.

3|1|Updated Mar 23, 2026
One-click install
npx skills add https://github.com/malinskibeniamin/skills --skill setup-sandcastle
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: setup-sandcastle
Source: https://github.com/malinskibeniamin/skills/tree/main/setup-sandcastle
Command: npx skills add https://github.com/malinskibeniamin/skills --skill setup-sandcastle

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Sandcastle orchestrates multiple autonomous agents across sandboxed environments to run tasks in parallel, enabling scalable delegation and HITL review workflows.

Core Features & Use Cases

  • Parallel agent execution across isolated sandboxes for GitHub issue work.
  • Headless run() for CI/batch automation and interactive() for human-in-the-loop reviews.
  • NoSandbox() mode for git worktrees and configurable branch strategies (head, merge-to-head, branch).

Quick Start

Install Sandcastle tooling, configure the environment, and run the main script to start parallel agent work.

Frequently Asked Questions about setup-sandcastle

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

FAQPage Schema
How do I run parallel autonomous agents across sandboxed environments for GitHub issues?

You can orchestrate parallel autonomous agents across isolated sandboxes using Sandcastle, which delegates GitHub issues into multiple agent threads for headless batch runs and human-in-the-loop reviews.

What is the best way to delegate GitHub issues into multiple agent threads for batch automation?

Delegating GitHub issues into multiple agent threads is best handled by Sandcastle's headless run() mode, enabling scalable parallel task execution and batch automation across sandboxed environments.

Does Sandcastle support human-in-the-loop reviews for autonomous agent workflows?

Yes, Sandcastle supports human-in-the-loop reviews through its interactive launch mode, allowing users to review and intervene in autonomous agent workflows during parallel task execution.

Do I need a Node.js environment to orchestrate agents with Sandcastle?

Yes, setting up Sandcastle for parallel agent orchestration requires a Node.js environment to run the main script and manage multiple autonomous agent threads across sandboxed environments.

Can I use git worktrees instead of isolated sandboxes for parallel agent execution?

Yes, Sandcastle provides a NoSandbox() mode that uses git worktrees instead of isolated sandboxes, supporting configurable branch strategies like head, merge-to-head, and branch for parallel tasks.

Why use configurable branch strategies when running autonomous agents in parallel?

Configurable branch strategies like head, merge-to-head, and branch are necessary to manage how parallel autonomous agents commit and integrate their work within sandboxed environments during task orchestration.