ThreeBrowserStudio AGENTS.md

Contribution rules for ThreeBrowser Studio, an editor where an AI agent authors 3D scenes through MCP and a native WebGPU window shows the live result.

In plain words
What is it for?
Use them when changing the 3D authoring kernel, MCP tools, runtime loading, scene data, release packaging, or generated runtime output.
Why use it?
They protect the product's lightweight design and keep the editable scene document, revisions, validation, and recovery behavior consistent.

Instructions file for CodexOpenCode

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.

agentmods
npx agentmods add instructions/samg-coder/threebrowserstudio/agents-md
Clone the repo
git clone --depth 1 https://github.com/SamG-Coder/ThreeBrowserStudio

Made for: Codex, OpenCode.

Per session 397 This file is loaded in full into every session.
When invoked 397 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
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 $0.00397 $0.00397
Opus 5 $0.00198 $0.00198
Sonnet 5 $0.00079 $0.00079
Haiku 4.5 $0.00040 $0.00040

Measured 2d ago against content hash 20a57462f54c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ThreeBrowserStudio 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 2d 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.

AGENTS.md · 32 lines

What it actually says

ThreeBrowser Studio contributor rules

ThreeBrowser Studio is an LLM-first authoring runtime. MCP is the editor and the native WebGPU window is the shared live viewport. Preserve that product shape: do not grow a conventional inspector-heavy browser UI.

  • Keep this repository standalone and lean. Never copy ThreeBrowser Runtime, Three.js, GTA, Secret River, or sample-game source/assets into it. The Windows release pack may copy already-built host binaries only; it must not vendor Runtime, ThreeC++, or sample source.
  • Resolve and launch the external ThreeBrowser Runtime through scripts/runtime-path.mjs; machine-local paths do not belong in projects.
  • The project-owned typed document is canonical. Three.js objects, UUIDs, generated JavaScript, node-material JSON, and RTX registrations are compiled runtime products only.
  • Every authored mutation goes through the core kernel with exact stable IDs, one revision, idempotency, validation, inverse history, and recovery.
  • Keep the nine MCP tools coarse and stable. Add typed operations or catalog entries before adding more top-level tools.
  • No raw eval, WGSL, GLSL, TSL, or unrestricted code through ordinary MCP mutations. User-authorized trusted project code remains a separate path.
  • Compile candidates before committing and swap visible results only after validation. Keep explicit GPU capture off the ordinary render loop.
  • Add dependencies only when the platform or a small local module cannot do the job. Do not add an app framework or bundle Three.js.
  • Run npm test after changes. For native rendering changes, also run npm run launch, exercise the live bridge, capture the exact window or GPU frame, and inspect the image before claiming completion.

AGENT_RULES.md contains the operating rules sent to models using Studio. DESIGN.md is the architectural contract and phased roadmap.

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. 2d ago First seen · 32 lines · 397 tokens per session scan A 20a57462f54c

Subscribe to this mod's changes

ThreeBrowserStudio AGENTS.md is an instructions file published in the GitHub repository SamG-Coder/ThreeBrowserStudio (17 stars, last pushed 2d ago), licensed MIT. It adds 397 tokens to every session, about $0.0020 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.

Related

Other instructions, from other repositories

Unity-AI-Animation CLAUDE.md

Instructions for IvanMurzak/Unity-AI-Animation, covering claude.md, what this is, build / run, critical invariants and find detail in.

IvanMurzak/Unity-AI-Animation · 377 tokens

blender-mcp AGENTS.md

Instructions for sandraschi/blender-mcp, covering blender-mcp agent context and quick ref.

sandraschi/blender-mcp · 79 tokens

blender-mcp CLAUDE.md

Instructions for sandraschi/blender-mcp, covering blender-mcp — ai-powered blender automation, commands, architecture and code style & standards.

sandraschi/blender-mcp · 239 tokens

blender-mcp copilot-instructions.md

Instructions for sandraschi/blender-mcp: You have access to a Blender 3D automation server. You can control Blender scenes, objects, materials, rendering, and physics through MCP tools.

sandraschi/blender-mcp · 85 tokens

HatchPet-CapybaraLulu AGENTS.md

Instructions for srwang0506/HatchPet-CapybaraLulu, a project described as: A lively Capybara Lulu desktop pet for Codex, with smooth expressive animation, task-aware reactions, 16-direction gaze, and complete source frames.

srwang0506/HatchPet-CapybaraLulu · 305 tokens

HatchPet-CapybaraLulu CLAUDE.md

Instructions for srwang0506/HatchPet-CapybaraLulu: Read AGENTS.md before modifying artwork, animation timing, installation behavior, or documentation. Claude Code can help maintain this repository, but it does not natively render Codex custom pets. Do not claim that copying this pet into /.claude creates a desktop…

srwang0506/HatchPet-CapybaraLulu · 66 tokens