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.
npx agentmods add skills/forgeyclap/claude-forge/forge-gamenpx skills add ForgeyClap/claude-forge --skill forge-gamegit clone --depth 1 https://github.com/ForgeyClap/claude-forgeWrote 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.
[](https://agentmods.dev/skills/forgeyclap/claude-forge/forge-game)<a href="https://agentmods.dev/skills/forgeyclap/claude-forge/forge-game"><img src="https://agentmods.dev/badge/skills/forgeyclap/claude-forge/forge-game.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00050 | $0.01393 |
| Opus 5 | $0.00025 | $0.00696 |
| Sonnet 5 | $0.00010 | $0.00279 |
| Haiku 4.5 | $0.00005 | $0.00139 |
Grade A, and why
forge-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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Forge playbook — Games (browser · Godot)
A game is an interactive-loop domain — "done" means a human (or a driven browser session) can actually play it start-to-finish, not that a screen renders. HUD/menus and responsive layout defer to forge-website / ui-boss; this file governs the playable loop, stuck-state safety, input/viewport coverage, and honest playtest evidence. The core honesty rule: never call a game playable without real evidence of a completed loop (win and lose and restart reached). Opt-in limit: browser games can be driven + screenshotted directly with the browser tool; Godot needs the Godot toolchain (editor/export), which is opt-in and owner-side — if it isn't installed, mark export/clean-run steps not-run and report only what was actually driven.
Hard rules (non-negotiable)
- A genuinely playable loop. The core loop runs end-to-end: input → state update → render → feedback, reaching a win and a lose condition and a working restart. Not a static mockup, not a title screen with a dead "Play" button.
- No stuck / soft-lock states. No dead-end where the player can't progress or restart; game-over and win are always reachable and recoverable; pause/resume works; input never stops responding. A soft-lock is a shipping-blocker, not a polish item.
- Cross-viewport + declared inputs. The canvas scales/letterboxes correctly at mobile and desktop sizes with no cut-off UI — or the game explicitly declares desktop-only. Every declared input method actually works (keyboard, mouse, touch, gamepad as appropriate; Godot: multiple resolutions/aspect ratios via a stretch mode).
- Honor the dependency constraint. If the brief says vanilla / no-framework / zero-dependency, honor it — no CDN pulls, self-contained assets. If the game ships as a Forge artifact, it must be self-contained under the artifact CSP (inline JS/CSS, assets as data URIs — no external hosts).
- Runtime health. Zero console errors during a full play session; a stable frame rate using a delta-time loop (
requestAnimationFramewith elapsed time; a fixed timestep for physics) rather than assuming 60fps; and cleanup on restart/scene change — listeners, intervals, timers, and audio nodes are removed so replaying doesn't leak or double-fire. - Persistence works if promised. If save/load or high-scores are in scope, they actually persist and restore correctly across reloads.
- Real playtest evidence. The loop is proven by actually playing it (or driving it via browser automation) with screenshots/recording showing win + lose + restart — evidence, never an assertion.
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.
- 5d ago First seen · 40 lines · 50 tokens per session scan A a68db3edf601
forge-game is a skill published in the GitHub repository ForgeyClap/claude-forge (2 stars, last pushed 1mo ago), licensed MIT. It adds 50 tokens to every session and 1,393 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 2026-08-31.
Other skills, from other repositories
avalon-8-10p
Eight-to-ten-player Avalon rules for a permanent Agent host, with the user choosing to join the randomized player pool or spectate.
flame-harness-evaluator
Phase 6 — skeptical QA. Run the game, watch it, then judge against the contract. Default = functional check; --strict adds quality and edge-case passes.
flame-harness-design
Phase 3 — define the Flutter designtokens.dart spec (palette, typography, spacing), the game's art/visual concept, and the asset/audio sourcing plan.
flame-harness-retro
Phase 11 — score the completed pipeline against Anthropic's 9 harness principles plus game quality, and write the retrospective.
flame-harness-contract
Phase 4 — propose verifiable completion criteria and mandatory hard gates; reach AGREED (1-pass default, multi-round negotiation in --strict).
flame-harness-plan
Phase 2 — write the game PRD in the project's defaultlanguage (core loop, mechanics, content metrics, win/lose, scope guard), map the lib/ structure, and assign app name, slug, and bundle id.