Crabbox is a remote execution control plane that lets a local command run on managed cloud capacity, an SSH host, or another delegated sandbox while syncing code, streaming output, and recording evidence. It is for maintainers, contributors, teams, AI agents, and automation that need repeatable remote testing, builds, browser checks, or platform-specific validation. The catalogue add-ons provide commands and instructions for controlling Crabbox runs.
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.
npx skills add openclaw/crabbox --skill crabbox-quickstartgit clone --depth 1 https://github.com/openclaw/crabboxWrote 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.
[](https://agentmods.dev/skills/openclaw/crabbox/crabbox-quickstart)<a href="https://agentmods.dev/skills/openclaw/crabbox/crabbox-quickstart"><img src="https://agentmods.dev/badge/skills/openclaw/crabbox/crabbox-quickstart/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.
<a href="https://agentmods.dev/skills/openclaw/crabbox/crabbox-quickstart"><img src="https://agentmods.dev/badge/skills/openclaw/crabbox/crabbox-quickstart.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00106 | $0.01575 |
| Opus 5.5 | $0.00042 | $0.00630 |
| Sonnet 5.5 | $0.00021 | $0.00315 |
| Haiku 4.5 | $0.00011 | $0.00158 |
Grade B, and why
crabbox-quickstart scanned grade B with 2 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 1mo 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.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
sudo apt-get install -y make Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Ubuntu with `git`, `curl`, `tar`, `python3`, `rsync`, and passwordless `sudo`. How it starts
The opening of the file, as written. The whole thing — 155 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Crabbox Quickstart
Crabbox runs your repository's commands on a disposable box — a container on your own machine, or a remote machine you lease. It syncs your working tree, runs one command, streams the output back, and exits with that command's code.
The loop
lease -> sync -> run -> read output -> stop
- Lease a box from a provider.
--provider local-containeruses Docker or Podman on your own machine: no account, no login, no spend. - Sync your current checkout onto it. Crabbox builds the file list from Git, so the directory must be a repository.
- Run one command there. Its exit code becomes Crabbox's exit code.
- Read the streamed output, and optionally timing and test-result files.
- Stop the box. A one-shot
rundoes this for you; a warm box does not.
Install and check
crabbox --version # already installed? skip the install
brew install openclaw/tap/crabbox # or github.com/openclaw/crabbox/releases
crabbox doctor --provider local-container
doctor with a provider is the readiness check that matters: it names the
container runtime, the leases you already hold, and the image= your commands
will run inside.
First run, no account needed
With Docker or Podman running, this works in any Git repository with no config file, no login, and no cloud spend:
git init # only if not already a repository
crabbox run --provider local-container -- uname -a
Crabbox provisioned a container, synced the dirty checkout, ran the command there, streamed its output, propagated its exit code, and deleted the lease. Budget 30-45 seconds once the base image is local, nearly all of it container startup; the very first run adds a one-time image pull.
The box is bare
Read this before swapping uname -a for real work. The default image is plain
Ubuntu with git, curl, tar, python3, rsync, and passwordless sudo.
No node, npm, make, gcc, go, cargo, or java. A command needing a runtime fails
before your code does — usually exit 127 and make: not found, or for npm a
preflight that stops the run first. Install what you need once, on a warm box:
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.
- 1mo ago First seen · 155 lines · 106 tokens per session scan B df3dca131592
crabbox-quickstart is a skill published in the GitHub repository openclaw/crabbox (1,481 stars, last pushed yesterday), licensed MIT. It adds 106 tokens to every session and 1,575 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it B with 2 findings (asks for root, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-13.
Other skills, from other repositories
agent-sandbox
Agent skill for sandbox - invoke with $agent-sandbox.
verify
Build the computer-use-demo image and drive the Streamlit UI headlessly to verify changes end-to-end.
test-locally
Build and deploy a full local OpenMetadata stack with Docker to test your connector in the UI. Handles code generation, build optimization, health checks, and guided testing.
redamon-testing
How RedAmon tests actually run and how to author them: the per-file Docker gate, the unit/integration/live tiers, and the failure modes that make a green run a lie. Trigger: editing any test.py, .test.ts(x) or tests/.sh; a test that is red, skipped or xfailed; a request to "run the tests", "make it green" or check…
bdd-container-update
Update the BDD test container image when its dependencies or runtime change. Determines version bump (major vs minor vs patch), rebuilds the Docker image, pushes to GitLab registry, and updates the image tag in starttest.sh. Test source (tests/, features/, scripts/, data/, conftest.py) is bind-mounted from the host…
testcontainers-integration-tests
Write integration tests using TestContainers for .NET with xUnit. Covers infrastructure testing with real databases, message queues, and caches in Docker containers instead of mocks.