Live survey304 models · 14 providers · 23,592 auditsAudited implement · SAFE →What changedHow we verifyllms.txt

MCP server · Audited from source

g-mcp-godot-studio

by hi-godot/godot-aiMIT licenceAudited 13 Nov 2026 · source 2e4d5529ca85

PyPI godot-aissestdiostreamable-http★ 3,523
✕BLOCKGrade D · 69/100 · 13 Nov 2026 · 2e4d552

Is it safe? No — BLOCK, grade D (69/100). It exposes 21 tools: 11 read, 10 write and 0 destructive. The audit recorded 35 findings (0 critical, 2 high, 3 medium, 27 low). Maintainer's description: Production-grade MCP server and AI tools for the Godot engine. A Snap to install. Totally free and fun.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “g-mcp-godot-studio (hi-godot/godot-ai)”. Retrieved 13 Nov 2026. https://modelsatlas.com/mcp/hi-godot/godot-ai

Why this is BLOCK. The audit found 0 critical and 2 high findings; the most severe is Network egress at src/godot_ai/attach/ensure.py:546. Each finding is listed below with its file and line.

Why this page is still public. People search for this server by name. Hiding it would leave them with the README's word alone; here they see what the code does. We list it so you can decide, and we do not show connect commands for BLOCK servers.

Trust audit

Audit 0.4.1 · full source read 13 Nov 2026

Trust score69/100✕BLOCK D
0 critical2 high3 medium27 low3 info

35 findings, listed below with the file and line each was found on. Grade A needs a sandboxed behavioural run too; this audit read the source only.

What the source does

  • FilesystemDeclared5 observations
  • NetworkDeclared11 observations
  • ShellNone observedNothing seen in the source
  • DependenciesPinnedNothing floats
  • SecretsFoundSee the findings

Findings (35)

Sorted by severity · the 25 most severe of 35 are listed

HIGHNetwork egressnet.env_exfil · src/godot_ai/attach/ensure.py:546
F-04 Network egressCWE-200 · CWE-319Layer L1
os.environ.copy() ... websocket
Why
reads secrets in the same file that sends data out
Fix
—
HIGHNetwork egressnet.tls_off · src/godot_ai/attach/ensure.py:345
F-04 Network egressCWE-200 · CWE-319Layer L1
timeout=timeout, trust_env=False, verify=False,
Why
certificate verification is disabled
Fix
leave verification on
MEDIUMNetwork egressnet.raw_ip · .claude/hooks/session-start.sh:88
F-04 Network egressCWE-200 · CWE-319Layer L1
http://127.0.0.1:8000/mcp; then
Why
No further detail recorded for this rule.
Fix
—
MEDIUMObfuscation / stealthobf.zero_width · script/_plugin_junction.ps1:1
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1
# Ensure test_project\addons\godot_ai is a directory junction to this
Why
No further detail recorded for this rule.
Fix
—
MEDIUMHard-coded secretssecret.generic · tests/integration/test_startup_call_timing.py:20
F-03 Hard-coded secretsCWE-798 · CWE-321Layer L1

Excerpt withheld: it is a secret, and we never print one.

Why
No further detail recorded for this rule.
Fix
—
LOWInventory / provenanceinv.binary · test_project/tests/fixtures/import_settings.wav
F-13 Inventory / provenanceCWE-1104Layer L0
import_settings.wav
Why
a compiled or binary member cannot be reviewed from source
Fix
ship source, or explain the binary in the README
LOWCode injectioncode.eval_exec · tests/unit/test_stormtest_auth_rotation.py:28
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
exec(compile(ast.Module(body=nodes, type_ignores=[]), "stormtest.py", "exec"), namespace)
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · tests/unit/test_stormtest_support.py:43
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
exec(compile(ast.Module(body=[function], type_ignores=[]), "stormtest.py", "exec"), namespace)
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · tests/unit/test_stormtest_support.py:650
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
exec(compile(ast.Module(body=functions, type_ignores=[]), "stormtest.py", "exec"), namespace)
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · tests/unit/test_stormtest_support.py:713
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
exec(compile(ast.Module(body=[function], type_ignores=[]), "stormtest.py", "exec"), namespace)
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWCode injectioncode.eval_exec · tests/unit/test_stormtest_support.py:770
F-02 Code injectionCWE-78 · CWE-94 · CWE-95Layer L1
exec(compile(ast.Module(body=[function], type_ignores=[]), "stormtest.py", "exec"), namespace)
Why
evaluates text as code
Fix
remove; use a parser or a dispatch table
LOWFilesystem / pathfs.traversal · script/qualification_index.py:58
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
f'<a href="../../files/{quote(row["filename"])}#sha256={row["sha256"]}"'
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · script/qualification_index.py:78
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
"url": "../../files/" + quote(row["filename"]),
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/unit/test_release_support.py:215
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
record["files"]["../../outside"] = {"size": 1, "sha256": "a" * 64}
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/unit/test_v4_release.py:458
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
attacker.writestr(f"{v4_release.PLUGIN_PREFIX}../../escape", b"attacker")
Why
No further detail recorded for this rule.
Fix
—
LOWFilesystem / pathfs.traversal · tests/unit/test_verify_worktree_link.py:90
F-05 Filesystem / pathCWE-22 · CWE-59Layer L1
link.symlink_to(Path("../../plugin/addons/godot_ai"))
Why
No further detail recorded for this rule.
Fix
—
LOWNetwork egressnet.env_exfil · tests/integration/test_attach_bridge.py:556
F-04 Network egressCWE-200 · CWE-319Layer L1
dict(os.environ) ... websocket
Why
reads secrets in the same file that sends data out
Fix
—
LOWNetwork egressnet.raw_ip · .github/workflows/nightly-diagnostics.yml:252
F-04 Network egressCWE-200 · CWE-319Layer L1
--target editor-a=http://127.0.0.1:8000/mcp \
Why
No further detail recorded for this rule.
Fix
—
LOWNetwork egressnet.raw_ip · CHANGELOG.md:716
F-04 Network egressCWE-200 · CWE-319Layer L1
`http://127.0.0.1:8000/mcp` entry can no longer authenticate. Use the dock's
Why
No further detail recorded for this rule.
Fix
—
LOWNetwork egressnet.raw_ip · README.md:64
F-04 Network egressCWE-200 · CWE-319Layer L1
`http://127.0.0.1:8000/mcp` entry cannot authenticate.
Why
No further detail recorded for this rule.
Fix
—
LOWNetwork egressnet.raw_ip · docs/STRESS_TESTING.md:105
F-04 Network egressCWE-200 · CWE-319Layer L1
SS_URL=http://127.0.0.1:8010/mcp python script/stormtest.py
Why
No further detail recorded for this rule.
Fix
—
LOWObfuscation / stealthobf.decode_call · src/godot_ai/handlers/editor.py:186
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1
image_bytes = base64.b64decode(img.get("image_base64", ""))
Why
No further detail recorded for this rule.
Fix
—
LOWObfuscation / stealthobf.decode_call · src/godot_ai/handlers/editor.py:248
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1
image_bytes = base64.b64decode(image_b64)
Why
No further detail recorded for this rule.
Fix
—
LOWObfuscation / stealthobf.decode_call · tests/unit/test_release_support.py:1136
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1
spki = base64.b64decode(b"".join(pem.splitlines()[1:-1]), validate=True)
Why
No further detail recorded for this rule.
Fix
—
LOWObfuscation / stealthobf.decode_call · tests/unit/test_v4_release.py:148
F-07 Obfuscation / stealthCWE-506 · CWE-94Layer L1
base64.b64decode("".join(v4_release.PUBLIC_KEY_PEM.splitlines()[1:-1]))
Why
No further detail recorded for this rule.
Fix
—

Tools it exposes

21 declared in source · 0 destructive

11 read10 write0 destructive
ToolWhat it doesRisk
animation_createCreate a new Animation clip inside an AnimationPlayerwrite
batch_executeExecute a list of editor sub-commands in order, stopping on first error.write
node_createCreate (spawn) a new node in the scene tree.write
node_set_propertySet a property on a node.write
project_runRun (play) the Godot project from the editor.write
scene_saveSave the currently edited scene to disk.write
script_createCreate a script file on disk: GDScript (.gd, validated) or C# (.cs, text only).write
script_patchAnchor-based string-replace edit on a .gd or .cs file.write
session_activateSet the active Godot editor session for subsequent tool calls.write
test_runRun GDScript test suites inside the connected Godot editor.write
editor_reload_pluginReload the Godot editor plugin.read
editor_screenshotCapture a screenshot of the Godot editor viewport or running game.read
editor_stateGet current Godot editor state: version, readiness, open scene, play state.read
logs_readRead recent log lines from the Godot editor, plugin, or running game.read
node_findFind nodes in the scene tree by name, type, or group.read
node_get_propertiesGet properties of a node.read
scene_get_hierarchyGet the scene tree hierarchy from the open scene.read
scene_openOpen an existing scene file (.tscn) in the editor.read
script_attachAttach a script to a node in the scene tree.read
wrappedreturn xread
wrapped_toolreturn x + 1read

Safer alternatives

Audit history

  1. ✕BLOCK D 69 / 100

    First audit · 2e4d5529ca85

README badge

Badges show the current verdict, BLOCK included

ModelsAtlas: BLOCK DMODELSATLASBLOCK · D

A badge cannot be hidden by the maintainer while the verdict is BLOCK. Fix the findings, push, and the re-audit updates it.

Questions about g-mcp-godot-studio

What is the g-mcp-godot-studio MCP server?
Production-grade MCP server and AI tools for the Godot engine. A Snap to install. Totally free and fun.
What tools does g-mcp-godot-studio expose?
21 in total: 11 read-only, 10 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is g-mcp-godot-studio safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (69/100) and found 2 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What credentials does g-mcp-godot-studio need?
No credential environment variables were found in its source, so it appears to need none.
How does g-mcp-godot-studio run?
It speaks sse, stdio and streamable-http, so it runs as a local process your client starts. It is published on PyPI as godot-ai.
How current is this page?
The grade is for one exact copy of the source (2e4d5529ca85), read on 2026-11-13. The repository is watched and re-audited when it changes.
I already installed g-mcp-godot-studio. What should I do?
Remove it from your client's MCP configuration, then rotate any API keys and tokens that were in your environment while it ran.
Can the verdict change?
Yes. The repository is watched and re-audited when its source changes. If the findings are removed the verdict updates, and the history on this page shows it.

Sources

  • hi-godot/godot-ai on GitHub — the source we read, its licence, stars and activityObserved · 13 Nov 2026
  • ModelsAtlas trust audit — BLOCK verdict for commit 2e4d5529ca85Observed · 13 Nov 2026

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.