openclaw/acpx is a headless command-line client for the Agent Client Protocol, a standard interface for communicating with compatible coding agents. Agents, orchestrators, and developers use it to manage persistent or one-off sessions, permissions, and machine-readable output across tools such as Codex, Claude Code, Gemini CLI, and OpenClaw, with the catalogue containing agents and instructions for those workflows.
Borrowing it
Nothing to install: this file belongs to openclaw/acpx. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/openclaw/acpx/main/.agents/skills/crabbox/SKILL.mdgit clone --depth 1 https://github.com/openclaw/acpxWrote 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/acpx/crabbox)<a href="https://agentmods.dev/skills/openclaw/acpx/crabbox"><img src="https://agentmods.dev/badge/skills/openclaw/acpx/crabbox/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/acpx/crabbox"><img src="https://agentmods.dev/badge/skills/openclaw/acpx/crabbox.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 | $0.00039 | $0.03303 |
| Opus 5.5 | $0.00016 | $0.01321 |
| Sonnet 5.5 | $0.00008 | $0.00661 |
| Haiku 4.5 | $0.00004 | $0.00330 |
Grade A, and why
crabbox 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 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.
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.
This is a copy
100% identical to crabbox — 845 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 367 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Crabbox
Remote and clean-machine proof. Packages. Docker. Live providers. Desktop. Cross-OS. The consumer repository owns when its validation needs a remote environment; Crabbox availability alone is not a reason to offload local work.
Backends:
blacksmith-testbox: trusted maintainer source. Prepared CI.tbx_....aws: direct brokered Crabbox. Fresh PRs. Custom sync/env/capture.cbx_....local-container: Docker fallback. Not remote proof.ssh: existing operator host. macOS/Windows/WSL2.
Always report provider, id, run URL, command, result. Never call Testbox “AWS Crabbox.”
Repository Contract
This canonical skill owns portable Crabbox policy and CLI operations only.
Consumer-specific setup belongs in that repository's AGENTS.md, package
scripts, hydration workflow, or another file outside the synchronized skill.
Resolve these placeholders from trusted repository instructions before running an example:
<check-command>: the repository's focused or broad validation command.<install-and-check-command>: its clean-container install plus validation.<trusted-bootstrap-script>: a maintainer-reviewed untrusted-PR bootstrap stored outside the untrusted checkout.<container-image>and<owner/repo#number>: the consumer's runtime and PR.
Never invent a missing command or copy a command from another consumer.
Authorization and Isolation
Routine use of the configured Crabbox/Testbox environment is part of completing the requested task; do not ask for separate approval. This includes creating, reusing and stopping task-owned leases, temporary state, and clean checkouts or worktrees needed for proof or a task-required Crabbox repair.
A dirty, missing or occupied checkout is a reason to use a clean task-owned checkout or worktree, not a permission blocker. Preserve existing checkouts, branches and unrelated edits. Keep source-trust, credential, production-access, budget and publication boundaries; routine-use approval does not waive them.
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 Changed · -345 lines · -4 tokens per session scan B → A f241f11e376a
- 2mo ago First seen · 712 lines · 43 tokens per session scan B ed512c0b0385
crabbox is a skill published in the GitHub repository openclaw/acpx (3,396 stars, last pushed 2d ago), licensed MIT. It adds 39 tokens to every session and 3,303 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. It is 100% identical to crabbox, differing in 845 lines, and is treated as a copy.
Other skills, from other repositories
verify-omnigent
Drive Omnigent the way a user does and prove a behavior with recorded evidence, using an isolated server, runner, and mock model plus a feature map of every user entry point. Load before reproducing a user-facing bug, before claiming a fix works, or when reviewing whether a change covered every surface (session vs.…
test-master
Generates test files, creates mocking strategies, analyzes code coverage, designs test architectures, and produces test plans and defect reports across functional, performance, and security testing disciplines. Use when writing unit tests, integration tests, or E2E tests; creating test strategies or automation…
eval-dataset-design
A guide to creating reliable test sets for AI agents. Each evaluation task includes the request, the simulated user's behaviour, a reset starting state, and a success check that can be independently verified.
dogfood
Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to dogfood, QA, exploratory test, find issues, bug hunt, or test this app on mobile.
cua-driver
Use Cua Driver for desktop or browser tasks that are awkward or unavailable through Bash/APIs, or when the user explicitly wants GUI interaction: app testing, visual bug reproduction, form filling, calendar entry, screenshots, and demo recording. Also covers Cua setup; not OpenAI Codex Computer Use or web research.
critique
Adversarial design critique of the current work. Render it, look at it, and argue for rejection. Run after the gates are green, never instead of them.