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 instructions/heagandev/threejs-agent-starter/agents-mdgit clone --depth 1 https://github.com/heagandev/threejs-agent-starterWrote 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/instructions/heagandev/threejs-agent-starter/agents-md)<a href="https://agentmods.dev/instructions/heagandev/threejs-agent-starter/agents-md"><img src="https://agentmods.dev/badge/instructions/heagandev/threejs-agent-starter/agents-md.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 | $0.00754 | $0.00754 |
| Opus 5 | $0.00377 | $0.00377 |
| Sonnet 5 | $0.00151 | $0.00151 |
| Haiku 4.5 | $0.00075 | $0.00075 |
Grade A, and why
threejs-agent-starter AGENTS.md 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 3d 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 — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Rules — Three.js Game Dev Starter Pack
You are helping a developer build or extend a browser game using this starter kit. Read this file fully before writing any code.
Stack
- Vite + TypeScript + Three.js (vanilla, no framework)
- All game logic lives in
src/main.ts - All styles live in
src/style.css - No additional packages unless explicitly requested
Architecture Rules
- Single class pattern — all game logic belongs inside one top-level game class named after the game (e.g.
AsteroidsGame,BlockBreakerGame). Do not create loose functions or module-level state outside the class. - Config object — all tunable values (speeds, sizes, positions, thresholds) go in
GAME_CONFIG. No magic numbers inline. - Game phases — use a const-object pattern (not TS
enum—erasableSyntaxOnlyforbids it):const GamePhase = { MENU: 'MENU', PLAYING: 'PLAYING', GAME_OVER: 'GAME_OVER' } as const. State is always one of these. Every behaviour must respect the current phase. - Typed interfaces — define TypeScript interfaces for all non-trivial objects (e.g.
Enemy,Projectile,Pickup). - Cleanup — any new event listeners, animation frames, or Three.js objects added must be disposed/removed in
destroy(). - No new files — unless explicitly asked, keep all code in
src/main.tsandsrc/style.css.
Three.js Rules
- Always cap pixel ratio:
Math.min(window.devicePixelRatio, 2) - Always cap delta:
delta = Math.min(delta, 0.04)to prevent physics explosions on tab switch - Use
THREE.Box3for AABB collision detection — update bounds each frame withsetFromObject() - Use
THREE.MathUtils.damp()for smooth deceleration, not manual lerp - Dispose geometries and materials when removing objects from the scene
- Shadow maps are enabled — new meshes that should cast/receive shadows must set those flags explicitly
Code Style
constoverletwherever possible- Private class fields with
readonlywhere applicable - Descriptive variable names — no single-letter variables except loop indices
- Group related logic into clearly named private methods
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.
- 3d ago First seen · 69 lines · 754 tokens per session scan A 57d4a2f3f95e
threejs-agent-starter AGENTS.md is an instructions file published in the GitHub repository heagandev/threejs-agent-starter (2 stars, last pushed 2mo ago), licensed MIT. It adds 754 tokens to every session, about $0.0038 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 instructions, from other repositories
kart-royale CLAUDE.md
Instructions for ryancampbell/kart-royale, covering working on this repo, commands, the contract, traps that have already cost a round and verifying a change.
Open-Party-Lab AGENTS.md
Instructions for Hartwich/Open-Party-Lab, covering ai agent guide, mental model, important entry points, public source boundaries and contribution rules.
3d-web-game-dev-skills AGENTS.md
Instructions for Shellishack/3d-web-game-dev-skills, covering repository instructions, summarization standard, generalization rules, skill content and review checklist.
webxr-first-steps copilot-instructions.md
Instructions for meta-quest/webxr-first-steps: This is a Meta Quest / Horizon OS sample.
vibe-coding-prompt-template backend.instructions.md
Instructions for KhazP/vibe-coding-prompt-template: Read AGENTS.md, agentdocs/techstack.md, and agentdocs/codepatterns.md.
claudemon CLAUDE.md
Instructions for zamarrowski/claudemon, covering ai agent guidelines, project shape, general guidelines, general rules and guards & defensive code.