Name: benchmark-sandbox
Rating: 78.4 (1 reviews)
Author: vercel

benchmark-sandbox

Run vercel-plugin eval scenarios in Vercel Sandboxes instead of local WezTerm panels. Provisions ephemeral microVMs with Claude Code + plugin pre-installed, runs benchmark prompts, extracts hook artifacts, and produces coverage reports.

2.09x

Quality

68%

Does it follow best practices?

Impact

92%

2.09x

Average score across 3 eval scenarios

Securityby

Medium

Suggest reviewing before use

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/benchmark-sandbox/SKILL.md

Quality

Content

70%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.

The content is exceptionally actionable with a clearly sequenced, validation-gated batch workflow and concrete code throughout. Its weaknesses are verbosity from duplicated flag/example sections and a monolithic structure with a referenced script that is not bundled.

Suggestions

De-duplicate the CLI usage: keep one flag table and one example block, then point 'Commands' back to it rather than restating the same invocations.

Move the dated 'Proven Results (2026-03-10)' metrics and the full JSON scoring schemas into a reference file (e.g. results.md / scoring-schemas.md) and link one level deep from SKILL.md.

Either bundle the referenced run-eval.ts under scripts/ or reframe the references so the SKILL.md does not depend on a script that is absent from the skill bundle.

Dimension	Reasoning	Score
Conciseness	At ~388 lines the body is noticeably verbose: the CLI flag table and bash examples are repeated across 'Proven Working Script' and 'Commands', and the dated 'Proven Results (2026-03-10)' section with version numbers is time-sensitive padding that penalizes token efficiency.	2 / 5
Actionability	Fully executable throughout — copy-paste bash invocations, exact CLI flags with defaults, TypeScript monitoring snippets, concrete JSON scoring schemas, and precise sandbox paths/commands cover the common cases.	5 / 5
Workflow Clarity	The 3-phase pipeline is clearly sequenced with explicit validation (haiku scoring after each phase), conditional gating ('VERIFY if >1 project file', 'DEPLOY if >3 project files'), and feedback loops (deploy retry up to 3x, verify fix-reverify), satisfying the batch-operation validation requirement.	5 / 5
Progressive Disclosure	Good section headers exist, but the SKILL.md is monolithic — JSON schemas, the ASCII session flow, proven-results tables, and the comparison table are all inlined rather than split into references, and the referenced run-eval.ts has no accompanying bundle directory.	3 / 5
	Total	15 / 20 Passed

Description

67%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.

The description is highly specific and distinctive, clearly conveying a multi-phase sandbox eval pipeline in third person. Its main weakness is the absence of an explicit 'Use when...' trigger clause, which leaves the activation condition only weakly implied.

Suggestions

Append an explicit trigger clause, e.g. 'Use when running vercel-plugin evals at scale, when you need parallel automated coverage with verification and deploy, or when local WezTerm panes are insufficient.'

Add a natural-language synonym or two ('sandbox benchmarks', 'remote plugin eval') so users searching with everyday phrasing match the skill.

Keep the existing concrete action list but lead with the trigger context so the 'when' is as prominent as the 'what'.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Provisions ephemeral microVMs with Claude Code + plugin pre-installed, runs benchmark prompts, extracts hook artifacts, and produces coverage reports' — covering the full pipeline comprehensively.	5 / 5
Completeness	The 'what' is clearly stated across all phases, but there is no 'Use when...' clause or equivalent explicit trigger guidance; the 'when' is only weakly implied via 'instead of local WezTerm panels', which per the guidelines caps completeness at 3.	3 / 5
Trigger Term Quality	Relevant domain keywords are present ('eval scenarios', 'Vercel Sandboxes', 'benchmark prompts', 'hook artifacts') but they are technical jargon with no natural-language trigger phrasing or synonyms a user would spontaneously say.	3 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (Vercel Sandbox-based eval) and explicitly contrasts with 'local WezTerm panels', giving it distinct triggers with minimal conflict risk against other skills.	5 / 5
	Total	16 / 20 Passed

Validation

100%

Warnings & errors only

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: vercel/vercel-plugin
Path: .claude/skills/benchmark-sandbox/SKILL.md
Commit: 11c3258

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

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.