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 skills add vucinatim/air-jam --skill airjam-mcpgit clone --depth 1 https://github.com/vucinatim/air-jamWrote 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/vucinatim/air-jam/airjam-mcp)<a href="https://agentmods.dev/skills/vucinatim/air-jam/airjam-mcp"><img src="https://agentmods.dev/badge/skills/vucinatim/air-jam/airjam-mcp/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.
<a href="https://agentmods.dev/skills/vucinatim/air-jam/airjam-mcp"><img src="https://agentmods.dev/badge/skills/vucinatim/air-jam/airjam-mcp.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00035 | $0.00702 |
| Opus 5 | $0.00017 | $0.00351 |
| Sonnet 5 | $0.00007 | $0.00140 |
| Haiku 4.5 | $0.00003 | $0.00070 |
Grade A, and why
airjam-mcp 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 11d 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 — 57 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Air Jam MCP
Use this skill when the project includes the Air Jam MCP and the task involves Air Jam-native dev workflows.
Default Order
- inspect the project through the Air Jam MCP first
- inspect the relevant game through the Air Jam MCP
- read canonical logs through the Air Jam MCP before adding custom logging
- run the smallest relevant quality gate through the Air Jam MCP
- fall back to shell commands only when the MCP does not expose the required operation
- when host staging actions exist, read the published action metadata before guessing payloads
Tool Bias
Prefer the Air Jam MCP for:
- project and game inspection
- unified dev log reads
- dev process start/stop/status
- topology inspection
- live game-session reads and host staging action invocation
- quality gates
For terminal-only agents, use the canonical pnpm exec airjam equivalents.
The persistent airjam session open/read/input/invoke/close lifecycle exposes
the same semantic game contract as MCP without requiring a client plugin and
returns JSON for composition.
If the game does not yet expose the control an agent needs, prefer adding a small game-owned host staging action over automating the visible browser UI.
If the template ships src/game/contracts/agent.ts, treat that file as the canonical semantic agent surface and use the MCP agent tools before inferring raw controller/store semantics.
When local UI verification matters, pair the Air Jam MCP with the current agent or client's built-in browser or in-app browser tooling first.
If Claude Code Desktop preview launches the app, use the committed .claude/launch.json; it runs Air Jam's preview-managed adapter for Claude Preview. If another browser or preview tool wants a launch command, use pnpm run dev.
Do not use raw vite for normal Air Jam development because it skips the local Air Jam backend.
If you need HTTPS/secure dev, use the normal pnpm run dev -- --secure flow.
For Claude Code or similar preview surfaces, keep host and preview controllers in the same visible preview screen when possible. Add multiple preview-controller windows from the host workspace for multiple local players instead of opening unrelated OS browser tabs unless the user asks for tabs.
For visible controller UI smoke proof, use real browser click/drag/release gestures on preview controllers. Do not synthesize pointer events into controller iframes from parent-page eval.
After editing host-only runtime refs, physics loops, or useHostActionListener side effects, hard refresh the host or run pnpm exec airjam reset local if actions appear duplicated or rendered state no longer matches replicated state.
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.
- 11d ago First seen · 57 lines · 35 tokens per session scan A 0113513d8e04
airjam-mcp is a skill published in the GitHub repository vucinatim/air-jam (6 stars, last pushed 2d ago), licensed MIT. It adds 35 tokens to every session and 702 once invoked, about $0.0002 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
design-lenses
Structured game-design critique: review a game through Schell's design lenses, Wright's systems-and-emergence lenses, and the MDA framework, producing severity-ranked findings with concrete fixes. Use when the user asks 'is my game fun?', 'review my game design', 'why is it boring?', 'critique this game', 'what's…
teach-me
Teach the user end-to-end browser game development over multiple sessions — design, engine, art, audio, juice, multiplayer, shipping — by building real games together. Use when the user wants to LEARN to make games themselves: 'teach me game development', 'I want to learn to make games', 'teach me Phaser', 'how do…
gaming-expert
Build comprehensive game systems including gameplay mechanics, multiplayer infrastructure, analytics tracking, and monetization features. Use when the user mentions game development, Unity or Unreal, gameplay mechanics, multiplayer netcode, game analytics, or in-game monetization.
game-design
Game design principles. GDD structure, balancing, player psychology, progression.
pc-games
PC and console game development principles. Engine selection, platform features, optimization strategies.
game-development
Game development orchestrator. Routes to platform-specific skills based on project needs.