claude

An integration that lets Claude Code run commands inside agentkernel sandboxes, which are isolated environments for code execution. It can add a sandbox skill and a /sandbox command to a project.

In plain words
What is it for?
Use it when Claude needs to run Python or other commands in an isolated sandbox, either automatically through the installed skill or explicitly with /sandbox.
Why use it?
It separates code execution from the main environment, helping limit the effects of commands while Claude works on a project.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/thrashr888/agentkernel/claude
Clone the repo
git clone --depth 1 https://github.com/thrashr888/agentkernel
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,070 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00000 $0.01070
Opus 5 $0.00000 $0.00535
Sonnet 5 $0.00000 $0.00214
Haiku 4.5 $0.00000 $0.00107

Measured 2d ago against content hash 02ff6a15acd3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

claude 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 2d 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.

docs/agents/claude.md · 162 lines

How it starts

The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Claude Code

Use Claude Code with agentkernel for isolated code execution.

Integration Levels

Level How It Works Setup
MCP Tools Claude calls agentkernel MCP tools for sandbox control agentkernel plugin install claude
Full Isolation Run Claude Code itself inside an agentkernel sandbox agentkernel sandbox create + agentkernel attach

Claude Code has its own native sandbox using OS-level primitives (Seatbelt on macOS, bubblewrap on Linux). That sandbox cannot be replaced with a custom runtime. The MCP plugin gives Claude access to agentkernel sandboxes as tools it can call.

Claude runs locally, code execution is sandboxed via MCP:

# Install the plugin into your project
agentkernel plugin install claude

# This creates:
#   .claude/skills/agentkernel/SKILL.md   — teaches Claude when/how to use sandboxes
#   .claude/commands/sandbox.md           — adds /sandbox command
#   .mcp.json                             — registers the agentkernel MCP server

Once installed, Claude can use sandboxed execution automatically via the skill, or you can invoke it explicitly:

/sandbox python3 -c "print('Hello from sandbox!')"
/sandbox npm test
/sandbox --profile restrictive python3 untrusted.py

For global installation (available in all projects):

agentkernel plugin install claude --global

Sandbox Mode

Run Claude Code itself inside an isolated sandbox:

# Create sandbox with Claude Code pre-installed
agentkernel sandbox create claude-dev --config examples/agents/claude-code/agentkernel.toml

# Start the sandbox
agentkernel sandbox start claude-dev

# Run Claude with your API key
agentkernel attach claude-dev -e ANTHROPIC_API_KEY=$ANTHROPIC_API_KEY

# Inside the sandbox:
claude

API Key

Claude Code requires an Anthropic API key. Get one from console.anthropic.com.

Read the full file on GitHub · 162 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. 2d ago First seen · 162 lines · 0 tokens per session scan A 02ff6a15acd3

Subscribe to this mod's changes

claude is an agent published in the GitHub repository thrashr888/agentkernel (58 stars, last pushed 3d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,070 tokens. 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 2026-08-30.