run-benchmark

run-benchmark is a skill for Claude Code from tmuskal/arc-agi-benchmarker. It costs 21 tokens per session (6,903 once invoked), scanned B, original, no licence file.

A skill for running benchmark tests on ARC-AGI games, a set of tasks used to evaluate whether an AI can solve unfamiliar visual reasoning problems.

In plain words
What is it for?
Having Claude Code play ARC-AGI games as an agent and recording the resulting scores.
Why use it?
It automates game runs and score recording, making repeated evaluation easier to compare.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the arc-agi-benchmarker plugin — 6 skills shipped together

Good fit Having Claude Code play ARC-AGI games as an agent and recording the resulting scores.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/tmuskal/arc-agi-benchmarker/run-benchmark
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 tmuskal/arc-agi-benchmarker --skill run-benchmark
Clone the repo
git clone --depth 1 https://github.com/tmuskal/arc-agi-benchmarker

Made for: Claude Code.

Or install arc-agi-benchmarker, the plugin that ships this one along with the rest of its 6 skills.

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 run-benchmark

README.md
[![agentmods](https://agentmods.dev/badge/skills/tmuskal/arc-agi-benchmarker/run-benchmark/github.svg)](https://agentmods.dev/skills/tmuskal/arc-agi-benchmarker/run-benchmark)
Your own site
<a href="https://agentmods.dev/skills/tmuskal/arc-agi-benchmarker/run-benchmark"><img src="https://agentmods.dev/badge/skills/tmuskal/arc-agi-benchmarker/run-benchmark/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.

agentmods 80×15 button for run-benchmark

Your own site · 80×15
<a href="https://agentmods.dev/skills/tmuskal/arc-agi-benchmarker/run-benchmark"><img src="https://agentmods.dev/badge/skills/tmuskal/arc-agi-benchmarker/run-benchmark.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 21 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 6,903 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 unknown 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 $0.00021 $0.06903
Opus 5 $0.00010 $0.03452
Sonnet 5 $0.00004 $0.01381
Haiku 4.5 $0.00002 $0.00690

Measured 12d ago against content hash 80b558402507, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade B, and why

run-benchmark 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 12d ago.

The scan reads SKILL.md. This mod also ships 4 executable files (scripts/enrich_meta.py, scripts/env_scores.py, scripts/finalize.py, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

find ~/.claude/plugins -path "*/arc-agi-benchmarker/*/scripts/game_driver.py" -exec cp {} .arc-agi-benchmarks/runs/<RUN_ID>/game_driver.py \; 2>/dev/null
plugins/arc-agi-benchmarker/skills/run-benchmark/SKILL.md · 598 lines

The source is not reproduced here

No licence file

A repository with no LICENSE is all rights reserved by default, so the body is not copied here. The metadata, the measurements and the link are.

Read it on GitHub

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 12d ago First seen · 598 lines · 21 tokens per session scan B 80b558402507

Subscribe to this mod's changes

run-benchmark is a skill published in the GitHub repository tmuskal/arc-agi-benchmarker (3 stars, last pushed 4mo ago), with no licence file. It adds 21 tokens to every session and 6,903 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

onboarding-and-teaching

Use when designing, reviewing, or generating the introduction to ANY mechanic in ANY genre — the four-beat teaching pattern (introduce→test→combine→twist), show-don't-tell level design, progressive disclosure and scaffolding, the mechanic dependency graph, FTUE (first-time user experience), early retention funnels…

rondorkerin/gamestack · 180 tokens

game-qa-lead

Invoke when the user asks about testing strategy, bug tracking, playtest methodology, regression testing, release certification, QA automation, or bug triage. Triggers on: "testing", "bug", "QA", "regression", "certification", "playtest methodology", "test plan", "release gate". Do NOT invoke for usability testing…

AlterLab-IEU/AlterLab_GameForge · 103 tokens

roblox-mobile-playtest

Playtest and QA Roblox experiences for mobile and tablet. Use for device simulator checks, touch controls, safe areas, thumb zones, text readability, orientation, HUD overlap, first-action friction, and mobile performance verification through Studio MCP.

AshExplained/roblox-skills · 52 tokens

roblox-playtest-qa

Playtest Roblox experiences through Studio MCP with attention to first-session fun, mobile UX, console errors, progression, social hooks, performance, and exploit-sensitive flows.

AshExplained/roblox-skills · 39 tokens

unity-agent-workflows

Use for AI-assisted Unity work that needs live repo discovery, project-derived routing, runtime-owner proof, runtime-visible output hard stops, runtime numeric proof for repeated visible-output failures, state-step guards, multi-agent scope ownership, modular C#/asmdef safety, UI/scene/visual asset gates, data-first…

hashgraph-online/awesome-codex-plugins · 146 tokens

roblox-studio-workflow

Modify living Roblox Studio projects safely: inspect Explorer and execution locations, preserve existing structure, reuse modules/remotes, create meaningful Instances, apply Attributes and CollectionService tags, test server/client and multiple devices, inspect Output, and remove experiments/debug objects. Use when an…

gamedev-skills/awesome-gamedev-agent-skills · 96 tokens