benchmark-testing

benchmark-testing is a skill for Claude Code from build-with-dhiraj/ai-workflow-framework-portability-kit. It costs 43 tokens per session (899 once invoked), scanned B, original, MIT.

A test-project launcher for checking how the Vercel plugin adds skills to Claude Code, an AI coding assistant. It creates separate projects and runs realistic prompts in WezTerm terminal panes.

In plain words
What is it for?
Creating benchmark projects, installing the plugin, launching Claude Code tests, and exercising skill injection across different scenarios.
Why use it?
It provides repeatable, isolated tests without manually preparing each project or terminal session.

Skill for Claude Code

Written for Claude Code: installed under .claude/. Also seen: reads .claude/ paths; mentions Claude Code.

Part of the vercel-plugin plugin — 7 skills, 2 agents, 1 MCP server shipped together

Good fit Creating benchmark projects, installing the plugin, launching Claude Code tests, and exercising skill injection across different scenarios.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/build-with-dhiraj/ai-workflow-framework-portability-kit/benchmark-testing
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add build-with-dhiraj/ai-workflow-framework-portability-kit --skill benchmark-testing
Clone the repo
git clone --depth 1 https://github.com/build-with-dhiraj/ai-workflow-framework-portability-kit

Made for: Claude Code.

Or install vercel-plugin, the plugin that ships this one along with the rest of its 7 skills, 2 agents, 1 MCP server.

Wrote 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.

agentmods badge for benchmark-testing

README.md
[![agentmods](https://agentmods.dev/badge/skills/build-with-dhiraj/ai-workflow-framework-portability-kit/benchmark-testing.svg)](https://agentmods.dev/skills/build-with-dhiraj/ai-workflow-framework-portability-kit/benchmark-testing)
Your own site
<a href="https://agentmods.dev/skills/build-with-dhiraj/ai-workflow-framework-portability-kit/benchmark-testing"><img src="https://agentmods.dev/badge/skills/build-with-dhiraj/ai-workflow-framework-portability-kit/benchmark-testing.svg" alt="Measured on agentmods" height="20"></a>
Per session 43 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 899 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00043 $0.00899
Opus 5 $0.00022 $0.00449
Sonnet 5 $0.00009 $0.00180
Haiku 4.5 $0.00004 $0.00090

Measured 7d ago against content hash 9d5ead7eed4f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade B, and why

benchmark-testing scanned grade B with 1 finding 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 7d 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.

Recursive force deletemediumDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

rm -rf ~/dev/vercel-plugin-testing

Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.

Plugins/vercel-marketplace-source/.claude/skills/benchmark-testing/SKILL.md · 73 lines

How it starts

The opening of the file, as written. The whole thing — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Benchmark Testing

Create isolated test projects that exercise vercel-plugin skill injection with realistic, technology-agnostic prompts.

Workflow

1. Create test directories

BASE=~/dev/vercel-plugin-testing
mkdir -p "$BASE"/{01-slug,02-slug,...}

2. Install the plugin in each directory

for dir in "$BASE"/*/; do
  echo "=== $(basename "$dir") ==="
  cd "$dir" && npx add-plugin https://github.com/vercel/vercel-plugin -s project -y 2>&1 | tail -1
done

This creates .claude/settings.json with enabledPlugins in each directory.

3. Launch Claude Code in WezTerm panes

Critical details that must all be followed:

  • Use --cwd <absolute-path> to set the working directory
  • Use unset CLAUDECODE before x to avoid nested-session errors
  • Use --settings .claude/settings.json (not --settings project) to load the plugin
  • Use double quotes on the outer -ic string, single quotes around the prompt
  • Wait 10 seconds between each launch to avoid overwhelming the system
  • Always use spawn (new tabs) — split-pane runs out of space after ~4 panes

Working command template:

wezterm cli spawn --cwd /absolute/path/to/test-dir -- /bin/zsh -ic "unset CLAUDECODE; x 'YOUR PROMPT HERE. Link the project to my vercel-labs team so we can deploy it later.' --settings .claude/settings.json; exec zsh"

Prompt Guidelines

  • Never name specific technologies (no "Next.js", "Stripe", "Vercel KV", etc.)
  • Describe the product and features — let the plugin infer which skills to inject
  • Make prompts ambitious and multi-featured to exercise multiple skill triggers
  • Always append: "Link the project to my vercel-labs team so we can deploy it later."

Example prompts

Slug Prompt Expected skills
recipe-platform "Build a recipe sharing platform where users sign up, upload photos of their dishes, write ingredients and steps, and browse a feed with infinite scroll..." auth, vercel-storage, nextjs
trivia-game "Create a multiplayer trivia game where players join a room with a 6-letter code, answer questions in real-time with a 15-second countdown..." vercel-storage, nextjs
code-review-bot "Build an AI-powered code review dashboard with webhook API routes, LLM streaming analysis, and stats over time..." ai-sdk, nextjs
conference-tickets "Create a conference ticketing system with tiered checkout, QR code emails, admin panel, and payment webhook handling..." payments, email, auth
content-aggregator "Build a content aggregator with hourly scheduled RSS fetching, LLM summaries, category filters, and bookmarks..." cron-jobs, ai-sdk
finance-tracker "Build a personal finance tracker with bank connection, spending charts, and weekly email digest via scheduled job..." cron-jobs, email
multi-tenant-blog "Create a multi-tenant blog where each user gets a subdomain, with request-level routing, headless content API, and role-based auth..." routing-middleware, cms, auth
status-page "Build a SaaS status page with scheduled endpoint pinging, uptime charts, incident logging, and KV-stored history..." cron-jobs, vercel-storage, observability
dog-walking-saas "Build a dog walking SaaS with user accounts, pet photos, booking, monthly invoicing, admin dashboard, and separate dev/prod env configs..." payments, auth, vercel-storage, env-vars

Read the full file on GitHub · 73 lines

Changes

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.

  1. 7d ago First seen · 73 lines · 43 tokens per session scan B 9d5ead7eed4f

Subscribe to this mod's changes

benchmark-testing is a skill published in the GitHub repository build-with-dhiraj/ai-workflow-framework-portability-kit (4 stars, last pushed 25d ago), licensed MIT. It adds 43 tokens to every session and 899 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it B with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.