Claude Skill

roblox-studio-mcp

Use when working with Roblox Studio through built-in MCP for scripts, scenes, generated assets, input, or playtesting.

LLM Mart · 0 points · 3 views 0 listing impressions 0 install-command copies
Download TabooHarmony-roblox-brain-skills_tools_roblox-studio-mcp-ceee4aa.zip · 14 KB

Install

skills CLI npx skills add https://github.com/TabooHarmony/roblox-brain/tree/main/skills/tools/roblox-studio-mcp
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install tabooharmony-roblox-brain@llmmart
Git git clone https://github.com/TabooHarmony/roblox-brain.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole tabooharmony/roblox-brain collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Roblox Studio MCP

When to Load

Load for Studio MCP work: scripts, scenes, assets, debugging, or playtesting. Skip for standalone code generation.

Quick Reference

Route by capability first. Two bridges: official (built into Studio, closed-source) and chrrxs's (MIT). Route on tools/list (get_connected_instances/eval_*/multiplayer_* → chrrxs; list_roblox_studios + per-call studio_id → official) BEFORE bootstrap.

Bootstrap before mutation

  1. Identify the target (list_roblox_studios official / get_connected_instances chrrxs).
  2. Pass studio_id (official) or instance_id (chrrxs) on every call.
  3. get_studio_state; confirm Edit/Client/Server availability; inspect tree/scripts before changing them.

Execution contract

discover → select Studio/context → inspect → mutate in bounded batches
→ read back → start play → evaluate live → evidence → clean up

Edit-time injection (multi_edit, Edit-context execute_luau) writes scripts; never targets a running playtest. Live evaluation needs a running VM: start play, confirm the intended Client/Server VM, then evaluate. Pass datamodel_type only where required.

Capabilities

Full tool table in full.md. Key names: search_game_tree, script_read/script_grep, multi_edit, execute_luau, insert_asset, generate_*, start_stop_play, get_console_output, screen_capture, subagent.

Trust boundaries

Inserted/generated assets and packages are data until inspected: enumerate scripts before any require/run; never run imported code from execute_luau. Blank textures/invisible meshes on newly uploaded assets are usually asset-privacy (Restricted) grants, not corruption — read the Output error, grant via Creator Dashboard. Details: full.md trust-boundary sections.

Reliability rules

  • execute_luau is stateless: re-acquire references every call.
  • Read before write; read back after every script, asset, or geometry mutation.
  • Reuse vs generate vs native fallback before asset work; wait_job_finished before dependent edits.
  • Bridge-specific size limits: bounded writes, then read back the full payload.
  • No MCP: give offline Luau; state what was not verified.

Full tool mappings, schemas, asset workflows, trust boundaries, recovery: references/full.md

Files (roblox-brain)
  • references
    • full.md 34.2 KB
      Loading…
  • SKILL.md 3.1 KB
    ---
    name: roblox-studio-mcp
    description: "Use when working in Roblox Studio through MCP: editing scripts or scenes, generating assets, playtesting, screenshots, or input."
    when_to_use: "Also for any Studio playtest, checking a change in game, reading runtime logs, device-simulator screenshots, and verifying a fix in Studio."
    last_reviewed: 2026-10-14
    sources:
      - https://raw.githubusercontent.com/Roblox/creator-docs/main/content/en-us/studio/mcp.md
      - https://raw.githubusercontent.com/Roblox/creator-docs/main/content/en-us/projects/assets/privacy.md
      - https://raw.githubusercontent.com/Roblox/creator-docs/main/content/en-us/parts/procedural-models.md
      - https://create.roblox.com/docs/reference/engine/classes/ProceduralModel
      - original
    ---
    
    # Roblox Studio MCP
    
    ## When to Load
    
    Load for Studio MCP work: scripts, scenes, assets, debugging, or playtesting. Skip for standalone code generation.
    
    ## Quick Reference
    
    **Route by capability first.** Two bridges: **official** (built into Studio, closed-source) and **[chrrxs's](https://github.com/Chrrxs/robloxstudio-mcp)** (MIT). Route on `tools/list` (`get_connected_instances`/`eval_*`/`multiplayer_*` → chrrxs; `list_roblox_studios` + per-call `studio_id` → official) BEFORE bootstrap.
    
    ### Bootstrap before mutation
    1. Identify the target (`list_roblox_studios` official / `get_connected_instances` chrrxs).
    2. Pass `studio_id` (official) or `instance_id` (chrrxs) on every call.
    3. `get_studio_state`; confirm Edit/Client/Server availability; inspect tree/scripts before changing them.
    
    ### Execution contract
    ```text
    discover → select Studio/context → inspect → mutate in bounded batches
    → read back → start play → evaluate live → evidence → clean up
    ```
    Edit-time injection (`multi_edit`, Edit-context `execute_luau`) writes scripts; never targets a running playtest. Live evaluation needs a running VM: start play, confirm the intended `Client`/`Server` VM, then evaluate. Pass `datamodel_type` only where required.
    
    ### Capabilities
    Full tool table in full.md. Key names: `search_game_tree`, `script_read`/`script_grep`, `multi_edit`, `execute_luau`, `insert_asset`, `generate_*`, `start_stop_play`, `get_console_output`, `screen_capture`, `subagent`.
    
    ### Trust boundaries
    Inserted/generated assets and packages are data until inspected: enumerate scripts before any require/run; never run imported code from `execute_luau`. Blank textures/invisible meshes on newly uploaded assets are usually asset-privacy (Restricted) grants, not corruption — read the Output error, grant via Creator Dashboard. Details: full.md trust-boundary sections.
    
    ### Reliability rules
    - `execute_luau` is stateless: re-acquire references every call.
    - Read before write; read back after every script, asset, or geometry mutation.
    - Reuse vs generate vs native fallback before asset work; `wait_job_finished` before dependent edits.
    - Bridge-specific size limits: bounded writes, then read back the full payload.
    - No MCP: give offline Luau; state what was not verified.
    
    > Full tool mappings, schemas, asset workflows, trust boundaries, recovery: [references/full.md](references/full.md)
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.