threebrowser-studio-mcp

A skill for building and editing Three.js WebGPU projects inside ThreeBrowser Studio, a visual environment for creating 3D scenes. It requires using the studio’s provided tools and live project state.

In plain words
What is it for?
Use it to create or modify 3D scenes, materials, shaders, textures, animations, lighting, cameras, and related project settings.
Why use it?
It keeps scene changes aligned with the actual Studio project and its supported tools, making the work inspectable and testable.

Skill for Claude CodeCodex

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 skills/samg-coder/threebrowserstudio/threebrowser-studio-mcp
Any agent
npx skills add SamG-Coder/ThreeBrowserStudio --skill threebrowser-studio-mcp
Clone the repo
git clone --depth 1 https://github.com/SamG-Coder/ThreeBrowserStudio

Made for: Claude Code, Codex.

Per session 122 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,586 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00122 $0.02586
Opus 5 $0.00061 $0.01293
Sonnet 5 $0.00024 $0.00517
Haiku 4.5 $0.00012 $0.00259

Measured 2d ago against content hash 7e27bf8bf4d1, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

threebrowser-studio-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 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.

skills/threebrowser-studio-mcp/SKILL.md · 212 lines

How it starts

The opening of the file, as written. The whole thing — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ThreeBrowser Studio MCP

Use the native Studio viewport as the shared canvas and the nine three_studio_* MCP tools as the sole project-authoring interface. Read docs/ai/patterns.md for categorized MCP patterns. AGENT_RULES.md and live three_studio_status stay authoritative.

Non-negotiable authoring boundary

  • Make every project read or mutation through three_studio_status, three_studio_project, three_studio_inspect, three_studio_apply, three_studio_validate, three_studio_render, three_studio_play, three_studio_history, or three_studio_job.
  • Never author by running a tutorial module, operation-list generator, bridge client, studio-call helper, JavaScript scene builder, or recorded macro.
  • Do not translate a finished JavaScript scene into one MCP batch. Decide and author each stage from the live project state and the latest visual evidence.
  • Shell access may launch Studio, OBS, or tests. It must not read, write, or mutate a Studio project behind MCP.
  • Do not use raw eval, JavaScript, WGSL, or GLSL as a scene-authoring escape hatch. Use supported typed resources and the live graph catalog.

Connect and establish authority

  1. Ensure the native Studio runtime is open. If the user wants a recording, start and verify recording before the first project or authoring call.
  2. Call three_studio_status first. Treat its capabilities, limits, revision, project identity, and tool schemas as authoritative.
  3. If a requested capability is false or missing, state the boundary and build the best honest supported result. Never claim RTX, scripting, physics, export, or gameplay when status does not expose it.
  4. Create or open the intended project with three_studio_project. For a new live build, use a fresh meaningful project path so the viewport begins from an honest empty/starter state.
  5. Inspect the active scene, current revision, and only the bounded catalog or resources needed for the next decision.

Read the full file on GitHub · 212 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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 · 212 lines · 122 tokens per session scan A 7e27bf8bf4d1

Subscribe to this mod's changes

threebrowser-studio-mcp is a skill published in the GitHub repository SamG-Coder/ThreeBrowserStudio (17 stars, last pushed 2d ago), licensed MIT. It adds 122 tokens to every session and 2,586 once invoked, about $0.0006 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.