check-arc-game-discoverable

check-arc-game-discoverable is a skill for Claude Code, Codex from theredbluepill/arc-interactive. It costs 53 tokens per session (2,559 once invoked), scanned A, original, MIT.

A review guide for checking whether an ARC game can be understood through play, without tutorials or written rules. ARC is a benchmark where players infer goals and controls by observing the game and trying actions.

In plain words
What is it for?
Use it to review game stems for discoverability, shared human-and-AI observations, meaningful feedback, and a realistic path from experimentation to winning.
Why use it?
It identifies games that may be technically playable but unfair because the player cannot discover the goal, rules, or useful feedback from the available actions and observations.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions AGENTS.md.

Good fit Use it to review game stems for discoverability, shared human-and-AI observations, meaningful feedback, and a realistic path from experimentation to winning.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/theredbluepill/arc-interactive/check-arc-game-discoverable
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 check-arc-game-discoverable
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 check-arc-game-discoverable

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

Your own site · 80×15
<a href="https://agentmods.dev/skills/theredbluepill/arc-interactive/check-arc-game-discoverable"><img src="https://agentmods.dev/badge/skills/theredbluepill/arc-interactive/check-arc-game-discoverable.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 53 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,559 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 $0.00053 $0.02559
Opus 5 $0.00026 $0.01280
Sonnet 5 $0.00011 $0.00512
Haiku 4.5 $0.00005 $0.00256

Measured 10d ago against content hash 18f0ee6bba9c, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

check-arc-game-discoverable 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 10d 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/check-arc-game-discoverable/SKILL.md · 131 lines

How it starts

The opening of the file, as written. The whole thing — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Check ARC game discoverability by play (no prior)

Soul of this skill

Assume a participant (human or AI) who has never seen this task, has not been trained on it, and receives no tutorial text, no labeled rule sheet, and no external description of goals or what each action “means.”

That is by design. In benchmark-style play, goals, rules, and action semantics are meant to be hidden from the game taker at the start. Participants are expected to figure them out themselves during play—by trying actions, watching what changes, hitting win/lose, and iterating.

Solvable here means exactly that: solvable by discovery through play—not “everything was explained or drawn on the HUD up front.” A stem passes if a reasonable participant can converge on enough of the goal and mechanics to win from experience alone (same observation and action channel an agent gets). It fails if, even while playing, the information or feedback needed to win never becomes available, or the space is so flat that learning cannot get off the ground.

If a human can do that, it is the foundation for claiming an AI without task-specific prior could in principle do the same under the same interface: the task must be learnable from interaction, not from documentation.

Human first, AI co-required

Under ARC-style benchmarking, human-solvable (in this skill’s sense) and AI-solvable under the official interface are meant to co-exist:

  1. Human first — Primary gate: can a cold-start participant learn goal, critical state, and action effects through play (observation + legal actions + consequences)? Review that discovery path, not whether the author pasted rules on screen.

  2. AI co-requiredBenchmark intent: human-solvable ⇒ AI-solvable (in principle) under the same official observation and action interface. Passing the human bar implies the task is fair for an agent that sees that channel: no extra win-critical information for humans only, levels mechanically winnable under the game’s actual logic, and no reliance on prose outside the env. It does not guarantee a particular model will win—only that the spec is not rigged against an honest AI relative to the human.

Read the full file on GitHub · 131 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. 10d ago First seen · 131 lines · 53 tokens per session scan A 18f0ee6bba9c

Subscribe to this mod's changes

check-arc-game-discoverable is a skill published in the GitHub repository theredbluepill/arc-interactive (54 stars, last pushed 2mo ago), licensed MIT. It adds 53 tokens to every session and 2,559 once invoked, about $0.0003 per session on Opus 5. 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.