CtrlK
BlogDocsLog inGet started
Tessl Logo

setup-review-sandbox

One-time setup of the sandbox prerequisites used by the reproduce-issue skill and the reproduce-verifier agent — Docker, the isolation runtime (gVisor on Linux / Colima on macOS), healthy container networking, and the nx-review-sandbox toolchain image (built from the repo's mise.toml). Idempotent; re-run any time to verify or repair. Use when the user says "set up the review sandbox", "install the sandbox prereqs", "build the sandbox image", or a reproduce-issue preflight reports something MISSING.

75

Quality

94%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Medium

Suggest reviewing before use

SKILL.md
Quality
Evals
Security

Quality

Content

88%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A strong, actionable setup runbook: idempotent check-first steps, platform branching, an explicit veth recovery loop, and a verification smoke test. The main weakness is a padded justifying anecdote in section 4 and a smoke-test variable that requires manual editing.

Suggestions

Condense the section-4 war story ("which is exactly how an image predating the pnpm-store warming went unnoticed for two weeks, costing ~25 minutes...") to one sentence — the 'don't gate on existence, Docker's layer cache makes unconditional builds cheap' rationale survives without the narrative.

Make the smoke-test $RUNTIME explicit instead of a comment-only hint, e.g. a one-liner 'RUNTIME=--runtime=runsc on Linux; empty on macOS' so the command is copy-paste ready without editing.

Move the long gVisor apt-install block into the existing tools/review-sandbox/ script directory and reference it, keeping SKILL.md a lean overview and pushing the >40-line install snippet one level deep.

DimensionReasoningScore

Conciseness

Mostly lean commands with terse, load-bearing notes (e.g. the bash -l / mise PATH gotcha), but the section-4 narrative about the stale pnpm-store image (~25 minutes over two weeks) is padding that could be condensed to a single rationale sentence.

4 / 5

Actionability

Copy-paste-ready bash for every step, platform-branched (Linux/macOS), with check-first idempotent probes and the sudo steps handed to the user as exact commands rather than described.

5 / 5

Workflow Clarity

Five clearly sequenced sections each gated by a check-first probe, an explicit error-recovery feedback loop for the veth/BTF failure (modprobe → reboot → persist), and a final smoke-test verification step.

5 / 5

Progressive Disclosure

Well-organized into five labeled sections with no nested references and a self-contained single-file structure, but at ~90 lines it sits above the under-50-line simple-skill threshold that would auto-award a 5; the long gVisor apt-install block could live in the referenced tools/review-sandbox/ scripts.

4 / 5

Total

18

/

20

Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A high-quality description: third-person voice, concrete prerequisites, idempotency noted, and multiple natural trigger phrases covering both user requests and preflight-driven invocations. It clearly separates what the skill does from when to invoke it.

DimensionReasoningScore

Specificity

Lists multiple concrete setup targets — Docker, the isolation runtime (gVisor on Linux / Colima on macOS), container networking, and the nx-review-sandbox toolchain image — giving comprehensive coverage of what gets configured.

5 / 5

Completeness

Explicitly answers both what (one-time idempotent setup of the named prerequisites) and when (concrete "Use when..." trigger phrases), matching the anchor for explicit what-and-when with concrete triggers.

5 / 5

Trigger Term Quality

Includes several natural user utterances ("set up the review sandbox", "install the sandbox prereqs", "build the sandbox image") plus a preflight-driven trigger ("reproduce-issue preflight reports something MISSING") that a user would plausibly say.

5 / 5

Distinctiveness Conflict Risk

Occupies a clear niche (review-sandbox provisioning) with distinct triggers and named sibling skills (reproduce-issue / reproduce-verifier), minimizing overlap with other skills.

5 / 5

Total

20

/

20

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

Total

15

/

16

Passed

Repository
nrwl/nx
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.