Guide · Agent skill
How to use the Build Game Monster System agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Build Game Monster System skill teaches an AI agent to create, integrate, audit or refactor enemy creatures for Three.js and browser-based action games. It is for developers who want monsters that work as real gameplay assets, with rigs, attack timing and hit detection that the game controls, rather than meshes that only look good. After this guide you will know what the skill does, how to install it, how to ask for it in plain language, and what your project needs before it is useful.
Build Game Monster SystemMengTo/Skills · MIT Download the skill Quick start
- 1Download the ZIP. Get the build-game-monster-system ZIP from this page.
- 2Unzip into a skills folder. Extract it into ~/.claude/skills/ for all projects, or into your project's .claude/skills/ folder for one project.
- 3Install related skills. Add img2threejs, build-hybrid-game-assets, tune-enemy-ai and design-action-combat if your work needs them.
- 4Confirm your game has the prerequisites. Make sure you have a Three.js project with combat events, an enemy state machine and a collision API.
- 5Ask for the monster task in plain language. Describe the monster and whether you want it built, integrated, audited or refactored.
- 6Review the conformance report. Check the rig, attack timing, tests and performance measurements, and any documented exceptions.
What does the skill make the agent do?
The skill tells the agent to build monsters as gameplay-compatible runtime assets. It separates each monster into four layers: a MonsterDefinition with serializable authored facts such as stats, attacks and asset provenance; a MonsterRig with joints, sockets and collision shapes; a MonsterRuntime that tracks current state, timers and combat events; and a MonsterViewAdapter that drives poses, effects, audio and LOD from runtime state.
A key rule is that the animation never decides hits. Each attack must declare its startup, active, recovery, cooldown, contact shape and interrupt policy, and damage comes from authoritative action IDs and collision events.
- Requires every monster to support idle, investigate, pursue, reposition, windup, attack, recover, stagger and defeated states.
- Asks for deterministic test fixtures for each attack phase, interrupt, death, LOD switch and reset.
- Ends with a conformance report: fully conforms, conforms with documented exceptions, or needs a new archetype capability added to the shared contract.
Install the skill from the ZIP or the Claude apps
Download the ZIP for this skill from this page. Unzip it into your personal skills folder at ~/.claude/skills/ so it is available in all your projects, or into a project's .claude/skills/ folder so it applies only to that project.
You can also upload the skill in the Claude apps under Settings → Capabilities → Skills.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload
Check your project before you ask for monsters
The skill asks the agent to read its contract reference first and to check your real game content model, combat events, enemy state machine, collision API, camera distance and asset loader before extending anything. Your project should already have these pieces. If it does not, the agent is told to check first, and it may stall or invent them.
The skill also refers to related skills for full functionality: img2threejs, build-hybrid-game-assets, tune-enemy-ai and design-action-combat. Install these as well if you need image-driven reconstruction, the procedural-versus-imported decision, enemy AI, or combat timing.
- A Three.js web project
- A game with combat events, an enemy state machine and a collision API
- Related skills: img2threejs, build-hybrid-game-assets, tune-enemy-ai, design-action-combat
- An in-app browser for the render checks in the verification steps
How to trigger the skill with natural requests
You do not need a special command. Describe the job in plain language and mention monsters, enemies, rigs or combat. The skill's description lists the topics it covers, so requests that use those words are most likely to load it. Name the monster and the task, and point the agent at your existing code where you can.
Useful request types include building a new enemy, adding a missing attack state, auditing an existing enemy for floating parts or missing states, and checking that every monster in the game follows the same standard.
- "Build a rigged skeleton-style enemy for my Three.js action game with windup, attack and recover states."
- "Audit the goblin enemy in src/enemies for missing states and floating attachments."
- "Refactor the boss monster so attack timing comes from the combat system, not the animation."
- "Add LOD levels to the wolf monster and check socket positions stay put."
- "Check whether all our monsters follow the shared monster contract and report any exceptions."
Review the output against the standard
Expect the agent to add or check the monster definition, rig hierarchy, sockets, collision shapes and attack definitions. It should also write fixtures and tests for state coverage, attack phase boundaries, action ID stability, joint propagation and disposal, then render the monster at real encounter distance in the in-app browser.
The skill asks the agent to measure triangles, draw calls, textures, frame time and mobile pixel-ratio behavior, and to fail the asset for floating attachments, missing mandatory states, pose-derived damage, LOD socket drift or unbounded resource cost. Read the final report and any listed exceptions before accepting the work.
- Rig uses +Y up, +Z forward, meters, and a ground-contact root
- Every required state exists and attacks declare their timing
- Damage comes from collision events, not from the pose
- Performance figures are measured, not assumed
What people use it for
- Building a new enemy with a skeleton-style rig and windup, attack and recover states for a Three.js action game.
- Auditing existing enemies for floating attachments, missing states or unbounded performance costs.
- Moving attack timing out of animation clips and into the game's combat system.
- Adding LOD levels to a monster while keeping sockets and collision shapes stable.
- Checking that every monster in a game follows one shared rig, combat and testing standard.
Frequently asked questions
How do I install the build-game-monster-system skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. You can also upload it in the Claude apps under Settings → Capabilities → Skills.
Is the build-game-monster-system skill free?
The skill is published under the MIT licence, as listed on this page. Check the licence file in the source repository for the full terms before reuse.
What do I need to use this skill?
You need a Three.js web project and a game that already has combat events, an enemy state machine and a collision API. The skill also works best with the related skills img2threejs, build-hybrid-game-assets, tune-enemy-ai and design-action-combat, and it uses an in-app browser for render checks.
Does this skill run scripts or use the network?
According to this listing, the skill does not run scripts and does not use the network. It consists of instructions and reference files that the agent reads.
Can I use this skill for Unity or Unreal projects?
No. The skill is written specifically for Three.js and web games, and its conventions and APIs would not match Unity or Unreal.
What if my project has no enemy state machine yet?
The skill asks the agent to check for existing combat events, an enemy state machine and a collision API first. If they are missing, the agent may stall or invent them, so add those systems before relying on the skill.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you build Three.js or web action games with several enemies and want every monster to share one rig, combat and testing standard. It is also useful for auditing existing enemy assets for missing states or performance problems. Skip it for Unity or Unreal projects, for simple decorative models with no combat, or for projects that lack a combat event system and enemy state machine.
Download the skill