Name: benchmark-sandbox
Rating: 73.6 (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

61%

Does it follow best practices?

Impact

92%

2.09x

Average score across 3 eval scenarios

Securityby

Advisory

Suggest reviewing before use

Optimize this skill with Tessl

npx tessl skill review --optimize ./.claude/skills/benchmark-sandbox/SKILL.md

Quality

Content

55%

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

This skill is extremely thorough and actionable with excellent workflow clarity — the 3-phase pipeline with validation checkpoints and error recovery is well-documented. However, it is severely over-long and monolithic, containing historical results, comparison tables, extensive 'hard-won discoveries', and repeated information that bloats the token cost significantly. The content would benefit enormously from splitting into referenced sub-files and aggressive trimming of narrative/historical content.

Suggestions

Split into multiple files: move JSON schemas to SCHEMAS.md, environment facts and discoveries to ENVIRONMENT.md, proven results to RESULTS.md, and the comparison table to a separate reference — keep SKILL.md as a concise overview with references.

Remove or drastically condense the 'Proven Results' section, 'Key Discoveries (Hard-Won)' narrative framing, and the 'When to Use This vs benchmark-agents' comparison table — these are reference material, not operational instructions.

Deduplicate the 'How It Works' numbered list and the 'Sandbox Session Flow' ASCII diagram — they describe the same process twice in different formats.

Remove explanatory phrases like 'Hard-Won', 'the proven eval runner', and editorial commentary — assume Claude can evaluate the information without persuasion.

Dimension	Reasoning	Score
Conciseness	Extremely verbose at ~400+ lines. Contains extensive historical context ('Proven Results (2026-03-10)'), repeated information across sections (sandbox session flow duplicates 'How It Works'), explanations of discoveries Claude doesn't need narrated ('Hard-Won'), and a comparison table with another tool. Much of this could be condensed to 1/3 the length.	1 / 3
Actionability	Highly actionable with concrete CLI commands, exact flags, executable TypeScript monitoring snippets, specific JSON schemas for scoring, precise file paths, and copy-paste ready examples throughout. Every section provides specific, executable guidance.	3 / 3
Workflow Clarity	The 3-phase pipeline is clearly sequenced with explicit validation checkpoints (haiku scoring after each phase), error recovery (deploy retries up to 3x, verify fix-and-retest loops), conditional gates (>1 file for verify, >3 files for deploy), and a detailed ASCII flow diagram showing the complete sandbox session lifecycle.	3 / 3
Progressive Disclosure	Monolithic wall of text with no references to external files despite the content clearly warranting separation. The JSON schemas, environment facts table, proven results, known limitations, and detailed phase descriptions could all be split into referenced files. No bundle files are provided, and no external references exist — everything is inlined into one massive document.	1 / 3
	Total	8 / 12 Passed

Description

67%

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 excels at specificity and distinctiveness, clearly articulating a narrow, well-defined set of actions for a specific technology stack. However, it lacks an explicit 'Use when...' clause, which caps completeness at 2, and the trigger terms are heavily technical jargon that may not match how users naturally phrase requests.

Suggestions

Add an explicit 'Use when...' clause, e.g., 'Use when the user wants to run vercel-plugin evaluations in cloud sandboxes instead of locally, or mentions eval benchmarks, sandbox testing, or coverage reports for vercel plugins.'

Include more natural trigger term variations such as 'test', 'evaluate', 'sandbox testing', 'remote eval', or 'cloud-based evaluation' to improve discoverability.

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions: provisions ephemeral microVMs, runs benchmark prompts, extracts hook artifacts, and produces coverage reports. Very detailed about what it does.	3 / 3
Completeness	Clearly answers 'what does this do' with specific actions, but lacks an explicit 'Use when...' clause or equivalent trigger guidance. The when is only implied by the description of the domain context.	2 / 3
Trigger Term Quality	Contains some relevant keywords like 'vercel-plugin', 'eval scenarios', 'Vercel Sandboxes', 'benchmark prompts', 'coverage reports', but these are fairly technical/niche terms. Missing common variations a user might say (e.g., 'test', 'evaluation', 'sandbox testing'). The contrast with 'local WezTerm panels' is helpful but narrow.	2 / 3
Distinctiveness Conflict Risk	Highly specific niche: vercel-plugin eval scenarios in Vercel Sandboxes with microVMs. Very unlikely to conflict with other skills due to the narrow, well-defined domain and explicit technology stack references.	3 / 3
	Total	10 / 12 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 — 11 / 11 Passed

Validation for skill structure

No warnings or errors.

Repository: vercel/vercel-plugin
Commit: 5641bff

Reviewed: about 21 hours 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.