sandbox

sandbox is a skill for Claude Code from vinnie357/claude-skills. It costs 88 tokens per session (2,010 once invoked), scanned A, original, MIT.

An index that helps choose an isolation method for an AI coding agent. It routes you to options based on the operating system, whether Kubernetes is used, the required isolation level, and the deployment size.

In plain words
What is it for?
Use it to select between Docker-based microVMs, gVisor, Kata Containers, and Apple Container approaches for agent workloads.
Why use it?
It prevents choosing a sandbox setup that does not fit the host or workload. It distinguishes single-machine setups from Kubernetes deployments and points to the relevant follow-up guide.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: reads .claude/ paths; mentions Claude Code.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the agent-sandboxing plugin — 7 skills, 6 commands shipped together

Good fit Use it to select between Docker-based microVMs, gVisor, Kata Containers, and Apple Container approaches for agent workloads.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add vinnie357/claude-skills
Claude Code
/plugin install agent-sandboxing

Made for: Claude Code.

Or install agent-sandboxing, the plugin that ships this one along with the rest of its 7 skills, 6 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for sandbox

README.md
[![agentmods](https://agentmods.dev/badge/skills/vinnie357/claude-skills/sandbox/github.svg)](https://agentmods.dev/skills/vinnie357/claude-skills/sandbox)
Your own site
<a href="https://agentmods.dev/skills/vinnie357/claude-skills/sandbox"><img src="https://agentmods.dev/badge/skills/vinnie357/claude-skills/sandbox/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for sandbox

Your own site · 80×15
<a href="https://agentmods.dev/skills/vinnie357/claude-skills/sandbox"><img src="https://agentmods.dev/badge/skills/vinnie357/claude-skills/sandbox.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 88 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,010 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00088 $0.02010
Opus 5 $0.00044 $0.01005
Sonnet 5 $0.00018 $0.00402
Haiku 4.5 $0.00009 $0.00201

Measured 7d ago against content hash 4ee7a7f4279e, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

sandbox scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/tools/agent-sandboxing/skills/sandbox/SKILL.md · 99 lines

How it starts

The opening of the file, as written. The whole thing — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.

sandbox

Entry-point index for the agent-sandboxing plugin. Load this skill first when the task is "give me an isolated environment for an AI agent" and you don't yet know which substrate fits. The body routes to the right deeper skill via decision tree + one-line index.

The plugin covers the Kubernetes-orchestrated path (multi-pod, per-session lifecycle, controller-managed). For the single-host CLI path (one operator, one machine, Docker microVMs), the sibling sbx plugin is the right tool — same mental goal, different shape.

Decision tree

Need to sandbox an AI agent
│
├─ Single host, macOS, TRUSTED workload needing a Linux dev environment?
│   └─→ /core:container                   (container machine, 1.0.0+ — host home mounted; NOT agent isolation)
│
├─ Single host, no Kubernetes?
│   └─→ /sbx:sbx                          (Docker microVMs, sbx CLI)
│
└─ Kubernetes (kind / GKE / kina)?
    │
    ├─ Host OS = macOS Apple Silicon?
    │   └─→ /agent-sandboxing:kina-microvm      (Apple Container microVM nodes)
    │
    ├─ Host OS = Linux + want microVM isolation?
    │   └─→ /agent-sandboxing:kata-on-kind      (Kata Containers on kind)
    │
    ├─ Production GKE + microVM isolation?
    │   └─→ /agent-sandboxing:kata-on-gke       (self-managed Kata on Ubuntu N2)
    │
    └─ Production GKE + userspace kernel isolation?
        └─→ /agent-sandboxing:k8s-agent-sandbox (covers GKE managed Agent Sandbox = gVisor)

For every k8s path: also load /agent-sandboxing:k8s-agent-sandbox (controller + CRDs + lifecycle) and /agent-sandboxing:claude-code-on-sandbox (workload packaging).

Isolation tier comparison

Tier Mechanism Where it runs Plugin skill
Hardware microVM (KVM) Kata Containers Linux + nested virt kata-on-kind, kata-on-gke
Hardware microVM (Virtualization.framework) Apple Container, node-level macOS Apple Silicon kina-microvm
Hardware microVM (KVM/Docker) sbx CLI, single-host Linux + macOS /sbx:sbx (sibling plugin)
Hardware microVM, host-integrated Apple Container machine (1.0.0+) macOS Apple Silicon /core:container — dev convenience tier, NOT isolation (host home mounted)
Userspace kernel gVisor Anywhere (no nested virt needed) k8s-agent-sandbox (gVisor variant)
Process-level (none) runc default Anywhere Not recommended for untrusted agents

Read the full file on GitHub · 99 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 7d ago First seen · 99 lines · 88 tokens per session scan A 4ee7a7f4279e

Subscribe to this mod's changes

sandbox is a skill published in the GitHub repository vinnie357/claude-skills (25 stars, last pushed today), licensed MIT. It adds 88 tokens to every session and 2,010 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

local-ai-agents

Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…

microsoft/ai-agents-for-beginners · 200 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

insight-error-page

Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…

vercel/next.js · 83 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens

next-partial-prefetching-adoption

Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…

vercel/next.js · 103 tokens