sandbox-agent

Deploy and operate Sandbox Agent across local and cloud sandboxes.

18|7|Updated Jan 23, 2026
One-click install
npx skills add https://github.com/rivet-dev/skills --skill sandbox-agent
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sandbox-agent
Source: https://github.com/rivet-dev/skills/tree/main/skills/sandbox-agent
Command: npx skills add https://github.com/rivet-dev/skills --skill sandbox-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Sandbox-agent enables developers to deploy, configure, and orchestrate AI coding agents inside sandboxed environments with a single, universal API.

Core Features & Use Cases

  • Deploy and manage Sandbox Agent instances across local and cloud sandboxes.
  • Create and manage agent sessions, stream events, and support human-in-the-loop interactions for multiple agents (Claude Code, Codex, OpenCode, Amp).
  • Use with SDKs or HTTP APIs to build coding UIs, dashboards, or automated pipelines.

Quick Start

Install the skill and follow the Quickstart guide in the SKILL.md to set up the server and begin creating sessions.

Frequently Asked Questions about sandbox-agent

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

FAQPage Schema
How do I orchestrate AI coding agents in sandboxed environments?

You can orchestrate AI coding agents in sandboxed environments by deploying Sandbox Agent to manage server setup, environment configuration, and session creation across local and cloud instances via a universal API.

Does sandbox-agent support human-in-the-loop workflows for Claude Code and Codex?

Yes, sandbox-agent supports human-in-the-loop workflows for Claude Code, Codex, OpenCode, and Amp by enabling message exchange and event streaming during active agent sessions.

Can I stream events from AI agent sessions using an HTTP API?

You can stream events from AI agent sessions using HTTP APIs or SDKs to build coding UIs, dashboards, and automated pipelines that monitor real-time agent operations.

What is the best way to manage multiple AI agent sessions across local and cloud sandboxes?

The best way to manage multiple AI agent sessions across local and cloud sandboxes is using a single universal API that handles environment management, session creation, and message exchange.

Do I need specific dependencies to set up sandbox-agent for AI orchestration?

You do not need specific external dependencies to set up sandbox-agent, but you must configure a root SKILL.md frontmatter with name and description, and may use optional scripts and references folders.