RhinoBLOCK
RhinoMCP connects Rhino 3D to AI Agent through the Model Context Protocol (MCP)
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Control Rhino 3D and Grasshopper with AI, in plain language.
RhinoMCP connects Rhino to AI agents through the Model Context Protocol, so assistants like Claude and Cursor can model geometry, read your document, and build Grasshopper definitions for you, just by chatting.
[](https://pypi.org/project/rhinomcp/) [](https://www.rhino3d.com/) [](https://www.python.org/) [](https://modelcontextprotocol.io/) [](LICENSE)
Quick start · What it can do · Usage · Examples · Tool reference
English · 简体中文
Highlights
For development of the autonomous improvement experiments, start with the continuation guide.
- Describe what you want and the assistant builds it in Rhino.
- It reads your document and can capture the viewport, so it works from what is actually on screen.
- It scripts Grasshopper for you: finding components, wiring them, setting sliders, and solving.
- A single plugin and a single config entry cover both Rhino and Grasshopper.
- When you need more control, it can run native Rhino commands, RhinoScript-Python, or RhinoCommon C#.
[!NOTE] RhinoMCP targets Rhino 8 on Windows and macOS.
Demos
[](https://youtu.be/pi6dbqUuhI4)
Two-way interaction: the AI
fd9801107445OBSERVED · 2026-09-26Connect
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.
claude mcp add rhinomcp -- uvx rhinomcp
{
"mcpServers": {
"rhinomcp": {
"command": "uvx",
"args": [
"rhinomcp"
]
}
}
}Exposed tools (49)
32 read · 15 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
analyze_objects | read | Inspect validity, bounding dimensions, and volume for created objects. |
assembly_command | read | Call an allowed command after reading its schema. Failures are recorded. |
boolean_difference | read | Subtract closed-solid cutters from a base solid. On success deletes sources |
boolean_intersection | read | |
boolean_union | read | |
capture_viewport | read | |
create_layer | write | |
create_object | write | Create a primitive. BOX params: width (X), length (Y), height (Z). |
create_objects | write | |
create_planar_region | write | Create one planar Brep face from closed coplanar boundary curves. |
create_source | destructive | Create exactly one declared new file; never overwrite an existing file. |
delete_object | destructive | Delete one construction object by ID in the dedicated experiment document. |
describe_command | read | Read an allowed command schema. Available: create_layer, create_object, |
describe_modeling_command | read | Read the exact existing RhinoMCP parameter schema before using a command. |
extrude_curve | read | Extrude a profile by a world XYZ vector. Cap a closed planar profile to |
get_or_set_current_layer | write | |
get_reference_image | read | Read one calibrated reference image. No candidate or hidden model access. |
gh_add_component | write | Add a Grasshopper component to the active canvas. |
gh_build_graph | write | Create and wire a Grasshopper graph in one batched canvas operation. |
gh_capture_preview | read | Capture live Grasshopper preview in a Rhino viewport without baking. |
gh_connect_components | read | Connect one Grasshopper output parameter to one input parameter. |
gh_create_document | write | Create or activate a Grasshopper document for subsequent canvas commands. |
gh_disconnect_components | read | Disconnect one Grasshopper wire, or all sources from a target input. |
gh_expire_solution | read | Expire the whole Grasshopper solution or selected components. |
gh_layout_components | read | Lay out Grasshopper canvas objects using wires as a left-to-right graph. |
gh_mutate_graph | write | Batch edit, rewire, and verify an existing or new Grasshopper graph. |
gh_run_solution | write | Run a Grasshopper solution and report runtime warnings/errors. |
gh_set_parameter_value | write | Set a slider, toggle, panel, value list, or regular input parameter value. |
gh_update_component | write | Update basic Grasshopper object metadata, position, enabled state, or preview state. |
inspect_view | read | Inspect your current model: perspective, front, right, back, left, top. |
intersect_curves | read | |
join_surfaces | read | Join 2..12 existing surfaces using Rhino |
loft | read | |
modeling_command | read | Call an allowed existing modeling command using its exact parameter schema. |
modify_object | write | |
modify_objects | write | |
offset_curve | read | |
pipe | read | |
project_curve | read | |
read_source | read | Read an explicitly allowed candidate source file and its SHA-256. |
redo | read | |
replace_source | read | Replace an allowed source file only if it still matches the last read hash. |
rotate_object | read | Rotate about the CURRENT WORLD BOUNDING-BOX CENTER, using XYZ radians. |
select_objects | read | |
split_curve | read | |
sweep1 | read | Sweep profiles along one rail using RoadlikeTop framing. Optional planar |
translate_object | read | Translate an existing object by a world XYZ vector in millimeters. |
undo | read | |
update_object_attributes | write |
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 | PASS |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (2 observation(s))
- Network
- declared (4 observation(s))
- Shell
- declared (2 observation(s))
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (11)
exec(compile(Path(spec.origin).read_text(), spec.origin, "exec"), module.__dict__)
mod = importlib.import_module(f"{__name__}.tools.{info.name}")create_source, delete_object
finding["evidence"][0]["path"] = "../../outside"
"get_modeling_guidance", {"topic": "../../etc/passwd"}base64.b64decode(result["image_data"])
base64.b64decode(result["image_data"])
path.write_bytes(base64.b64decode(result["image_data"]))
(directory / "perspective.png").write_bytes(base64.b64decode(result["image_data"]))
base64.b64decode(content.get("data", ""), validate=True)server/src/rhinomcp/static/rhinoscriptsyntax.py
Gates applied: no_behavioural_pass.
fd9801107445full audit observations/trust-audit/mcp-server/jingcheng-chen__rhino-3.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 |
|---|---|---|---|---|---|
| 2026-09-26 | fd9801107445 | BLOCK | D | 69 | first audit |
Questions
What is the Rhino MCP server?
RhinoMCP connects Rhino 3D to AI Agent through the Model Context Protocol (MCP)
What tools does Rhino expose?
49 in total: 32 read-only, 15 that write, and 2 that can delete or overwrite (create_source, delete_object). Every one is listed on this page with its risk.
Is Rhino 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. Separately from the audit: 2 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Rhino need?
No credential environment variables were found in its source, so it appears to need none.
How does Rhino run?
It speaks stdio, so it runs as a local process your client starts. It is published on PyPI as rhinomcp.
How current is this page?
The grade is for one exact copy of the source (fd9801107445), read on 2026-09-26. The repository is watched and re-audited when it changes.