play-arc-game

play-arc-game is a skill for Claude Code, Codex from theredbluepill/arc-interactive. It costs 0 tokens per session (1,487 once invoked), scanned A, original, MIT.

A guide for playing and testing ARC-AGI-3 games with the project's game runner. ARC-AGI-3 is a benchmark of interactive games and puzzles.

In plain words
What is it for?
Use it to run a selected game for a fixed number of steps, play it with keyboard or mouse controls, resolve its full game identifier, and test it through Python.
Why use it?
It provides the exact commands and input modes needed to run a game randomly, from a terminal, or by hand, reducing confusion when testing a game.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to run a selected game for a fixed number of steps, play it with keyboard or mouse controls, resolve its full game identifier, and test it through Python.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/theredbluepill/arc-interactive/play-arc-game
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 theredbluepill/arc-interactive --skill play-arc-game
Clone the repo
git clone --depth 1 https://github.com/theredbluepill/arc-interactive

Made for: Claude Code, Codex.

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 play-arc-game

README.md
[![agentmods](https://agentmods.dev/badge/skills/theredbluepill/arc-interactive/play-arc-game/github.svg)](https://agentmods.dev/skills/theredbluepill/arc-interactive/play-arc-game)
Your own site
<a href="https://agentmods.dev/skills/theredbluepill/arc-interactive/play-arc-game"><img src="https://agentmods.dev/badge/skills/theredbluepill/arc-interactive/play-arc-game/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 play-arc-game

Your own site · 80×15
<a href="https://agentmods.dev/skills/theredbluepill/arc-interactive/play-arc-game"><img src="https://agentmods.dev/badge/skills/theredbluepill/arc-interactive/play-arc-game.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 0 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,487 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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 $0.00000 $0.01487
Opus 5 $0.00000 $0.00744
Sonnet 5 $0.00000 $0.00297
Haiku 4.5 $0.00000 $0.00149

Measured 9d ago against content hash 118b5ba04831, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

play-arc-game scanned grade A with 0 findings 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 9d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/play-arc-game/SKILL.md · 146 lines

How it starts

The opening of the file, as written. The whole thing — 146 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Play ARC Game Skill

Skill for playing and testing ARC-AGI-3 games using run_game.py.

Running Games

(<stem> is the two-letter+digits id from GAMES.md, e.g. ez01. --version auto picks the sole package dir under environment_files/<stem>/.)

Random-agent mode (default)

Omit --mode or pass --mode random-agent — runs --steps random picks among ACTION1–ACTION5 (default 100 steps).

uv run python run_game.py --game <stem> --version auto --mode random-agent --steps 50

Terminal mode (typed 1–7)

uv run python run_game.py --game <stem> --version auto --mode terminal

Hand-play (pygame window)

Add --mode human — opens scripts/human_play_pygame.py (WASD / arrows, click for ACTION6 in display space). There is no separate matplotlib player in this repo.

Programmatic Testing

arc.make needs the full game_id string from that package’s metadata.json (e.g. ez01-63be02fb). From repo root you can resolve it with scripts/env_resolve.full_game_id_for_stem("ez01").

from arc_agi import Arcade, OperationMode
from arcengine import GameAction

arc = Arcade("environment_files", OperationMode.OFFLINE)
env = arc.make("ez01-63be02fb", seed=0)  # replace with your tree’s metadata game_id

# Execute actions
result = env.step(GameAction.ACTION1, reasoning={"test": 0})
print(f"State: {result.state}")

Other stems: pb02 (two-crate push), nw01 (arrow-tile forcing), ex01 (uses ACTION5 on the exit pad), gp01 / lo01 (ACTION6-only play; ACTION1–4 are no-ops) — each uses its own game_id from disk.

ACTION6: Click/Coordinate Actions

For games using ACTION6 (click), coordinates are in display space (0-63), not grid space.

Display to Grid Conversion

For a 32×32 grid on a 64×64 camera:

# Grid to display (what you send in action.data)
display_x = grid_x * 2 + 1
display_y = grid_y * 2 + 1

# Display to grid (what you receive)
grid_x = (display_x - 1) // 2
grid_y = (display_y - 1) // 2

Read the full file on GitHub · 146 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. 9d ago First seen · 146 lines · 0 tokens per session scan A 118b5ba04831

Subscribe to this mod's changes

play-arc-game is a skill published in the GitHub repository theredbluepill/arc-interactive (54 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,487 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.