Roblox Studio MCPBLOCK
MCP server for Roblox Studio runtime debugging, playtest control, screenshots/input, multiplayer testing, and per-peer server/client eval from AI agents.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Connect your coding agent directly to Roblox Studio. It can edit places, run Luau in live server and client contexts, start and stop playtests, and collect logs, screenshots, memory reports, and profiler captures from each peer.
[](https://www.npmjs.com/package/@chrrxs/robloxstudio-mcp)
What it can do
Debug a running game
- Run Luau with
eval_server_runtimeoreval_client_runtime. Both tools execute in a live server or client context and use the samerequirecache as your game scripts. - Instrument live code with
breakpoints. It records each hit without pausing the playtest. - Read output from edit mode, the server, or a specific client with
get_runtime_logs, including messages logged during startup.
Automate playtests
- Start, inspect, and stop solo or multi-client sessions with
solo_playtestandmultiplayer_playtest. - Open or close Studio windows with
manage_instance. It can launch a baseplate, a local place file, a published place, or an older place revision.
Find performance problems
- Record server or client CPU timings with
capture_script_profilerandcapture_micro_profiler. - Break down memory use with
get_memory_breakdownor attribute scene cost withget_scene_analysis.
Work in edit mode
- Run Luau in Studio's edit context with
execute_luau. - Use
set_propertiesfor instance properties andfind_and_replace_in_scriptsfor script text. For project-specific bulk edits, useexecute_luau. - For large generated Luau, use the verified chunk-staging workflow: explicit instance routing, UTF-8 byte/hash readback, ownership-checked cleanup, and bounded recovery without blindly replaying mutations.
- Use
selectionto inspect or update Studio selection and frame a part or model before capturing the viewport. - Capture the viewport with
capture_screenshot, then send mouse or keyboard input. See
8953a18d2b0aOBSERVED · 2026-10-12Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add studio-plugin --env AUTH_EMPTY_READS=${AUTH_EMPTY_READS} --env AUTH_TEST_HOME=${AUTH_TEST_HOME} --env AUTH_TOKEN_FILE=${AUTH_TOKEN_FILE} --env ROBLOX_OPEN_CLOUD_API_KEY=${ROBLOX_OPEN_CLOUD_API_KEY} -- npx -y [email protected]{
"mcpServers": {
"studio-plugin": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"AUTH_EMPTY_READS": "${AUTH_EMPTY_READS}",
"AUTH_TEST_HOME": "${AUTH_TEST_HOME}",
"AUTH_TOKEN_FILE": "${AUTH_TOKEN_FILE}",
"ROBLOX_OPEN_CLOUD_API_KEY": "${ROBLOX_OPEN_CLOUD_API_KEY}"
}
}
}
}Exposed tools (3)
3 read · 0 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
Gems | read | |
VIP | read | Perks |
noop | read |
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- none-observed
- Shell
- declared (2 observation(s))
- Dependencies
- not all pinned
- Secrets in source
- found
Findings (25)
'.git-credentials', '.gitconfig', '.netrc', '_netrc', '.pypirc', '.dockercfg',
Baseplate.rbxl
apiKey: 'asset-delivery-only-key',
token: 'active-parent-test-lock',
payload-timeout-diagnostics.test.ts
luau-payload-transfers.mjs
studio-payload-boundaries.mjs
const hash = createHash('md5');const hash = createHash('md5');const md5 = (bytes) => createHash('md5').update(bytes).digest('hex');'config/token.json', '.git-credentials', '.netrc', '.aws/credentials',
'.ssh/id_ed25519', '.claude.json', '.codex/auth.json', 'mcp.json',
const wrapper = path.resolve(__dirname, '../../../', packageName, 'src/install-plugin.ts');
const source = readFileSync(resolve(__dirname, '../../../../studio-plugin/src/modules', file), 'utf8');
entryPoints: [path.resolve(process.cwd(), '../../studio-plugin/src/modules/handlers/MetadataHandlers.ts')],
entryPoints: [path.resolve(process.cwd(), '../../studio-plugin/src/modules/CooperativeJobRunner.ts')],
const source = readFileSync(resolve(__dirname, '../../../../studio-plugin/src/modules/handlers/PropertyHandlers.ts'), 'utf8');
new URL(`http://127.0.0.1:${address.port}/mcp`),baseUrl: `http://127.0.0.1:${(server.address() as AddressInfo).port}`,const baseUrl = `http://127.0.0.1:${address.port}`;const baseUrl = `http://127.0.0.1:${address.port}`;const EDIT_SERVER_URL = 'http://127.0.0.1:58741';
'粒子Еffесt\u200D',
@modelcontextprotocol/client, @modelcontextprotocol/node, @modelcontextprotocol/server, @typescript-eslint/eslint-plugin, @typescript-eslint/parser, eslint, proper-lockfile, saxes
fzstd, saxes, @modelcontextprotocol/client, @modelcontextprotocol/node, @modelcontextprotocol/server, @types/express, @types/jest, @types/node
Gates applied: no_behavioural_pass.
8953a18d2b0afull audit observations/trust-audit/mcp-server/chrrxs__robloxstudio-mcp.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-12 | 8953a18d2b0a | BLOCK | D | 69 | first audit |
Questions
What is the @chrrxs/robloxstudio-mcp MCP server?
MCP server for Roblox Studio runtime debugging, playtest control, screenshots/input, multiplayer testing, and per-peer server/client eval from AI agents.
What tools does @chrrxs/robloxstudio-mcp expose?
3 in total: 3 read-only, 0 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is @chrrxs/robloxstudio-mcp safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (69/100) and found 1 critical or high issue in the source. Each one is listed on this page with the file and line it is on.
What credentials does @chrrxs/robloxstudio-mcp need?
It reads AUTH_EMPTY_READS, AUTH_TEST_HOME, AUTH_TOKEN_FILE, ROBLOX_OPEN_CLOUD_API_KEY, ROBLOX_STUDIO_AUTH_TOKEN and ROBLOX_STUDIO_NO_AUTH from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does @chrrxs/robloxstudio-mcp run?
It speaks streamable-http, so it runs as a service you connect to over the network. It is published on npm as studio-plugin at 1.0.0.
How current is this page?
The grade is for one exact copy of the source (8953a18d2b0a), read on 2026-10-12. The repository is watched and re-audited when it changes.