gamedev-skills/awesome-gamedev-agent-skills / skills/other-engines/roblox-studio-workflow
roblox-studio-workflow
VerifiedSafely edits an existing Roblox place or Rojo project: finds the source of truth and proves changes with the right test mode or Studio MCP (execute_luau). Use when an agent edits a .rbxl or Rojo project, works through Studio MCP tools, or must report what was tested. Not for installing Studio MCP (`roblox-get-started`) or the feature code (`roblox-luau`, `roblox-ui`).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 9.2 KB |
| agents/openai.yaml | config | 222 B |
| references/studio-mcp.md | markdown | 3.9 KB |
| references/studio-verification.md | markdown | 4.8 KB |
references/studio-verification.md
View on GitHub# Roblox Studio verification evidence
Read this before claiming a change was tested in Studio.
## Select modes by risk
- **Run**: server/world behavior without a player.
- **Test / Test Here**: quick single-session player path or camera-position spawn; toggle the
Client/Server view while it runs.
- **Server & Clients**: separate server and up to eight clients; required for meaningful remote,
ownership, targeted replication, join/leave, and simultaneous-player checks.
- **Device Simulator**: resolutions, orientation, DPI and scaling, on-screen keyboard, touch, and
device switching mid-playtest. It does not simulate device memory or performance; test real
hardware for that.
- **Controller Emulator**: input mappings and gamepad UI navigation.
- **Network Simulator**: separate inbound and outbound latency, jitter and packet loss on playtest
connections, with presets for wired, Wi-Fi and mobile; use it for Server Authority prediction and
`UnreliableRemoteEvent` behavior. It affects Studio playtests only.
- The toolbar versions of Device Simulator and Network Simulator were a Studio beta on 2026-11-11
(File → Beta Features → *New Device Simulator*, then restart). If the user cannot find them, ask
whether that beta is on.
- **`TestService`**: Studio-only scripted assertions (`TestService:Check()`, `Require()`,
`Message()`, `Fail()`) surfaced in the Output/TestService pane. Use it where the installed
Studio/API and project tooling expose observable results; do not invent a headless workflow.
- **`StudioTestService`** (plugin context): `ExecuteMultiplayerTestAsync(numPlayers, args)`
starts a server with 1-8 client DataModels and yields until the server calls `EndTest(result)`;
`ExecutePlayModeAsync` and `ExecuteRunModeAsync` do the same for solo Test and Run. `AddPlayers`,
`LeaveTest` and `GetTestArgs` script join and leave flows (`GetTestArgs` has a known issue when
called from a client LocalScript). Pair it with `UserInputService:CreateVirtualInput()` for
simulated mouse and keyboard input (it refuses to touch CoreGui) and `StudioDeviceSimulatorService` for scripted device
profiles. All are Studio-only and do nothing in a published experience.
## Evidence record
For each run record Studio version/date, place/build source, mode, client count, emulated device or
controller, setup, action, expected result, observed result, relevant server/client Output, and
pass/fail. Screenshots help visual review but do not replace interaction or Output evidence.
When the Roblox Studio MCP server is connected, its `get_console_output` and `screen_capture`
tools supply this Output and screenshot evidence directly, and `start_stop_play` plus the
playtest input tools drive the run; the manual testing modes remain the default when it is not.
Separate verification levels explicitly:
- **Executed:** observed in a running Studio session with Output/result evidence.
- **Statically checked:** parsed, linted, type-checked, or inspected without engine execution.
- **Prepared, not run:** an exact scenario/harness exists but requires manual Studio interaction.
- **Unverified:** no adequate check was possible.
Never collapse the latter three into “Studio tested.”
## Temporary harness rules
Keep a harness isolated and easy to remove. It may expose a pure validator or create representative
instances, but must call the production boundary rather than reimplement it. Guard Studio-only
fixtures so they cannot run in a published server. Remove test remotes, Parts, scripts, tags,
Attributes, debug UI, and logs before final hierarchy review.
## UI quality scenario
For an inventory/HUD/settings evaluation, author a real `ScreenGui` hierarchy with header, primary
content, secondary detail, repeated-item template, and action/footer region. Populate empty, small,
and overflow datasets. Verify desktop, narrow portrait, mobile landscape, tablet-like, and console;
mouse hover/press, touch activation, gamepad focus and scroll-follow; safe insets; long text;
disabled/error state; open/close interruption; respawn/reset policy; and Explorer readability.
Reject and revise if visual hierarchy depends on universal rounded cards, random gradients/neon,
oversized headings, excessive empty space/transparency/blur, identical action hierarchy, or noisy
motion. Fix the skill guidance—not only the fixture—when the output followed an underspecified rule.
## Networking scenario
Use at least two clients for valid, malformed, spam, out-of-range, stale character, leave during
work, rapid respawn, simultaneous requests, targeted response, and broadcast. Check authoritative
state in server context and visibility in every client, plus limiter/cleanup state after leave.
## Primary reference
- `https://create.roblox.com/docs/studio/testing-modes`
- `https://create.roblox.com/docs/reference/engine/classes/StudioTestService`
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |