sandbox — security grade SAFE, quality 66/100

Security audit verdict: SAFE · quality 66/100

No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →

by cocoonstack · MCP Server · ★ 61

Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h

🔒 Is sandbox safe to install? View the security audit →

About sandbox

sandbox MicroVM sandboxes for AI agents, built on cocoon: a fast-boot guest stack, an in-guest product daemon, a per-node control plane with warm pools, and SDKs for Go and Python. Warm claims are sub-millisecond; a pool miss clones from a golden snapshot in tens of milliseconds; cold boot is 0.2–0.4 s on bare metal. Cloud Hypervisor serves both network lanes. has no NIC and uses vsock-only I/O (hardened default); attaches a bridge/CNI NIC. Documentation: cocoonstack.github.io/sandbox (deployment, clusters, HTTP API, Go + Python SDK references, the MCP server, the OpenAI Agents SDK and LangChain adapters, silkd protocol, performance) — source in . Layout — in-guest product daemon (Rust, tokio): exec with context, persistent shell sessions, streaming fs, tar-stream tree push/pull, find/replace, watch (ready-acked), pty, structured git, guest port relay (), and an LSP broker for flavor-shipped language servers — newline-JSON frames over vsock 2048, RPCs back to bac

aiai-agentsfastmcp-servermicrovmopenaisandboxvirtualization

Quick Facts

Stars61
Forks4
LanguageGo
CategoryMCP Server
LicenseAGPL-3.0
Quality Score65.5646330742229/100
Open Issues13
Last Updated2026-09-20
Created2026-07-02
Platformsgo, mcp
Est. Tokens~17k

Compatible Skills

These tools work well together with sandbox for enhanced workflows:

  • clampdown — semantic(0.28)+complementary+same_lang+similar_pop+shared_platform (60%)
  • gitmoot — semantic(0.17)+complementary+same_lang+similar_pop+shared_platform (56%)
  • greywall — semantic(0.17)+complementary+same_lang+similar_pop+shared_platform (56%)
  • clawker — semantic(0.15)+complementary+same_lang+similar_pop+shared_platform (55%)

sandbox alternative? Top 6 similar tools

Looking for a sandbox alternative? If you're comparing sandbox with other mcp server tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.

  • runtm by runtm-ai · ⭐ 296

    Open-source sandboxes where coding agents build and deploy. Spin up isolated environments where Claude Code, C

  • SecGPT by llm-platform-security · ⭐ 121

    An Execution Isolation Architecture for LLM-Based Agentic Systems

  • cordium by octelium · ⭐ 75

    Open-source, general-purpose sandbox platform for devs and AI agents that provides identity-based secure acces

  • brood-box by stacklok · ⭐ 71

    CLI tool for running coding agents inside hardware-isolated microVMs

  • oreilly-ai-agents by sinanuozdemir · ⭐ 299

    An introduction to the world of AI Agents

  • deepcontext-mcp by Wildcard-Official · ⭐ 275

    DeepContext is an MCP server that adds symbol-aware semantic search to Claude Code, Codex CLI, and other agent

More MCP Server Tools

Explore other popular mcp server tools:

View all MCP Server tools →

Popular Go Agent Tools

Frequently Asked Questions

What is sandbox?

sandbox is Fast cold-boot MicroVM sandboxes for AI agents on cocoon. It is categorized as a MCP Server with 61 GitHub stars.

What programming language is sandbox written in?

sandbox is primarily written in Go. It covers topics such as ai, ai-agents, fast.

How do I install or use sandbox?

You can find installation instructions and usage details in the sandbox GitHub repository at github.com/cocoonstack/sandbox. The project has 61 stars and 4 forks, indicating an active community.

What license does sandbox use?

sandbox is released under the AGPL-3.0 license, making it free to use and modify according to the license terms.

What are the best alternatives to sandbox?

The top alternatives to sandbox on Agent Skills Hub include runtm, SecGPT, cordium. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.

How this security grade is produced

Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.

The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.

Sources & who's responsible:

View on GitHub → Browse MCP Server tools