Borrowing it
Nothing to install: this file belongs to alfredoperez/speckit-companion. 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/alfredoperez/speckit-companion/main/.claude/commands/sandbox.mdgit clone --depth 1 https://github.com/alfredoperez/speckit-companionWrote 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/commands/alfredoperez/speckit-companion/sandbox)<a href="https://agentmods.dev/commands/alfredoperez/speckit-companion/sandbox"><img src="https://agentmods.dev/badge/commands/alfredoperez/speckit-companion/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.
<a href="https://agentmods.dev/commands/alfredoperez/speckit-companion/sandbox"><img src="https://agentmods.dev/badge/commands/alfredoperez/speckit-companion/sandbox.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 | $0.00022 | $0.00841 |
| Opus 5.5 | $0.00009 | $0.00336 |
| Sonnet 5.5 | $0.00004 | $0.00168 |
| Haiku 4.5 | $0.00002 | $0.00084 |
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 19d 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.
How it starts
The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Your task
Build one test sandbox, or show what there is. A sandbox is a small throwaway project in a known starting state; the recipes live in the sibling speckit-sandboxes repo and every built sandbox lands under its sandbox root, never in this repo.
Shell state does not persist between Bash calls, so start each one with . .claude/sandboxes-env.sh. It sets $SANDBOXES_REPO (the sibling checkout) and $SANDBOXES_DIR (the sandbox root). If it says there is no checkout, relay that line and stop.
Read $ARGUMENTS
-
empty: list. Print the recipes, then the sandboxes that already exist:
. .claude/sandboxes-env.sh && "$SANDBOXES_REPO"/new-sandbox.sh --list . .claude/sandboxes-env.sh && ls -1 "$SANDBOXES_DIR" 2>/dev/null || echo "(no sandboxes yet)"Show both as two short lists, with the root's path above the second. Do not build anything. Two folders in the second list are not sandboxes, so say what they are when they show up:
bench-cellsholds the benchmark's cells andawesome-copilotis/publish-canvas's upstream clone. -
where: print the sandbox root and nothing else.. .claude/sandboxes-env.sh && "$SANDBOXES_REPO"/new-sandbox.sh --where -
<recipe> [name]: build one. The name defaults to the recipe, and a name that is taken is never reused: the new sandbox becomes<name>-2, then-3.. .claude/sandboxes-env.sh && "$SANDBOXES_REPO"/new-sandbox.sh <recipe> [name] -
<recipe> [name] --dev: the same, with Companion installed from this checkout in place of the published build.COMPANION_DIRpoints the recipe here; without it the recipe uses thespeckit-companionfolder beside the sandboxes repo, which is not this checkout when you are in a worktree.. .claude/sandboxes-env.sh && COMPANION_DIR="$PWD" "$SANDBOXES_REPO"/new-sandbox.sh <recipe> [name] -- --devA recipe with no Companion (
vscode-qa-stock,copilot-stock,claude-mod-stock) refuses--dev; say so and build nothing.
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.
- 19d ago First seen · 55 lines · 22 tokens per session scan A ee227d5a898b
sandbox is a command published in the GitHub repository alfredoperez/speckit-companion (99 stars, last pushed 6d ago), licensed MIT. It adds 22 tokens to every session and 841 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-10-06.
Other commands, from other repositories
test-tools
Systematic smoke tests for Backseat Driver tools - validates core functionality and basic error handling after tool updates.
e2e-testing
:::info The Playwright suite launches the real Electron main process and built React UI on Windows and macOS. It exercises the preload bridge, IPC handlers, and file storage. See Testing Commands for the separate Vitest suite. ::.
test-write
Turn a SOLID-verdict ## Test plan critique into real, executing test code via kit:test-writer, one case per matrix row. Never dispatches against a missing, stale, or non-SOLID verdict.
tdd-refactor
Safe refactoring verified by existing tests.
eg-fix-bug
Fix a bug using the elephant/goldfish workflow — problem doc, goldfish diagnosis check, failing test, fix, review, validate.
test-factory
Quick testing helper for generated skills and agents to ensure they work correctly.