Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

io.github.KittyCAD/zoo-mcp

PYPI · ZOO_MCP

An MCP server that provides access to the Zoo API for various CAD operations and tools.

Available components

−2 this week 60 Trust /100

Last scanned (23 hours ago)

Scores higher than 53% of PyPI servers

Category Design and Media

Recent critical change

CVE-2026-112398 affects this package (24 Oct 2026). See the changelog before you install this server.

Trust breakdown (7 categories)

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. How we score → Why this is hard to score →

Supply Chain Security49
  • Malware scan not yet available for this package.Unverified
  • No known CVEs affecting this package version or its production dependencies.Pass
  • No install/post-install scripts declared.Pass
  • 9 of 70 dependencies flagged as unhealthy. View diagnostics → Partial
Provenance & Transparency45
Schema Quality & AI Usability62
  • AI-judged instruction clarity (good).Pass
  • Context-footprint check failed: tool/resource definitions use about 9680 tokens (~158/item across 61 items; 61 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
  • Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management80
  • Stability check failed: the tool surface changed between 0.28.13 and 0.37.0: 6 tool removals, 0 breaking changes, 21 additions. See how to fix → Fail
Tool Coverage72
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 1% of tool parameters carry a description.Partial
  • Structured output schemas are declared (74% of tools); any adoption earns full credit.Pass
Tool Safety75
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • 0 of 5 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "execute_kcl" implies "execute" and declares no destructiveHint at all, which the MCP spec reads as destructive by default. See how to fix → Fail
  • An AI judge read all 61 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
  • Implements a current MCP spec version (2026-07-28).Pass
What we tested

Checks we ran (15)

  • Known CVEs
  • Install scripts
  • Dependency health
  • Source repository
  • Provenance
  • License
  • Maintenance
  • Security disclosure
  • Schema quality
  • Stability
  • Tool coverage
  • Injection markers
  • Destructive annotations
  • Judged manipulation
  • MCP protocol

Checks we could not complete (1)

These are gaps on our side, not findings about this package. We only credit what we can confirm, so a category made up entirely of them scores 0.

  • Malware scan in Supply Chain SecurityMalware scan not yet available for this package.

Does not apply to this package (2)

  • Endpoint SecurityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.
  • Transport & ReachabilityScored on hosted endpoints only. A package has no live endpoint of its own for us to probe.

The score does not measure everything. What the score does not measure →

Changelog

Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.

  • 21 Nov 26 −1
    • Package version: 0.36.0 → 0.37.0 functional
  • 20 Nov 26 −16
    • Malware scan: pass → unverified ▼ security
    • Tool coverage: 92% → 78% ▼ functional
    • Package version: 0.35.0 → 0.36.0 functional
  • 19 Nov 26 +15
    • Malware scan: unverified → pass ▲ security
  • 18 Nov 26 −15
    • Malware scan: pass → unverified ▼ security
    • Schema quality: 195 → 175 ▲ functional
    • Schema quality: excellent → good functional
    • Package version: 0.34.4 → 0.35.0 functional
  • 17 Nov 26 +15
    • Malware scan: unverified → pass ▲ security
  • 16 Nov 26 −15
    • Malware scan: pass → unverified ▼ security
  • 15 Nov 26 +15
    • Malware scan: unverified → pass ▲ security
  • 13 Nov 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 84 to 87.

Diagnostics

Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.

Captured 21 Nov 2026 · Analysed pypi/zoo_mcp@0.37.0

Provenance No attestation

The registry publishes no build provenance for this version, so there is nothing to verify.

Result No attestation
Ecosystem pypi
Reason No attestation published

Background: How many MCP packages publish verified provenance →

Dependencies 70 packages
Packages resolved 70
Stale 5
No linked repository 4
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 61 exposed · ~9,680 tokens

The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →

Tool Tokens
search_kcl_samples ~233

Search KCL samples by keyword. Searches across all KCL sample titles and descriptions for the given query. Returns matching samples ranked by relevance. Args: query (str): The search query (case-insensitive). max_results (int): Maximum number of results to return (default: 5). Returns: list[dict]: List of search results, each containing: - name: The public project ID (use with get_kcl_sample) - title: Human-readable title - description: Brief description of the sample - multipleFiles: Whether the sample contains multiple KCL files. Best-effort hint only — see list_kcl_samples for the full caveat. Call get_kcl_sample if you need a reliable answer. - match_count: Number of times the query appears in title/description - excerpt: A relevant excerpt with the match in context Raises: ToolError: The sample index could not be fetched; the call can be retried.

NameTypeReqDescription
max_resultsinteger––
querystringyes–
NameTypeReqDescription
result–yes–

No examples provided.

search_org_dataset_semantic ~153

Semantic-search a dataset for samples relevant to the query. Embeds the query with the org-dataset embedding model and returns the top chunk matches ranked by cosine similarity. Args: dataset_id (str): The UUID of the dataset to search (from `list_org_datasets`). query (str): The natural-language query to embed and search with. limit (int | None): Optional max number of matches to return. Returns: A list of match dicts (source_file_path, content, similarity, chunk_index, conversion_id), or an error message if the operation fails.

NameTypeReqDescription
dataset_idstringyes–
limit–––
querystringyes–
NameTypeReqDescription
result–yes–

No examples provided.

select_entities ~299

Replace the "selection set" with the given entities. Takes UUIDs exactly as they appear in an artifact graph, and works for faces, edges, and solids alike. Selected entities render with a distinct color tint and an outline, which is considerably more obvious in a `snapshot` than the subtler brightening that highlight_set_entities applies. For the strongest emphasis on a face, call both this and highlight_set_entities; for an edge use just one, as selection overrides the highlight. Follow with `snapshot` (passing zoom=False to keep the current camera) to see the result. Selections draw through the solid. An entity facing away from the camera still renders tinted, seen through the body, so a tinted entity in a snapshot is not evidence that it faces the camera: an occluded edge shows up as an interior diagonal rather than on the silhouette, and an occluded face looks washed out rather than solid. Pick a camera on the same side as the entity to see it properly. Args: entity_ids: Entity UUIDs to select; pass an empty list to clear the selection. session_id: An open modeling session, from start_modeling_session. Required: the selection is scene state and would be discarded without one. Returns: SelectReplace: Confirmation that the selection was replaced.

NameTypeReqDescription
entity_idsarrayyes–
session_idstringyes–

Structured output declared, but exposes no named fields.

No examples provided.

set_selection_filter ~97

Set which model entity types can be added to the "selection set". Args: entity_types: Entity types permitted by the selection filter. session_id: An open modeling session, from start_modeling_session. Required: the filter is scene state and would be discarded without one. Returns: SetSelectionFilter: Confirmation that the filter was set.

NameTypeReqDescription
entity_typesarrayyes–
session_idstringyes–

Structured output declared, but exposes no named fields.

No examples provided.

snapshot ~917

Render an open modeling session as an image. Populate the scene first by passing this session_id to execute_kcl. The camera always uses an orthographic projection, so measurements read off the image are not distorted by perspective. Args: session_id: A modeling session populated by execute_kcl. camera_view: Which view or views to capture. Omit it for a single isometric view. Otherwise one of: 1. A named view: 'front', 'back', 'left', 'right', 'top', 'bottom', 'isometric', 'isometric_front_right', 'isometric_front_left', 'isometric_back_right', 'isometric_back_left'. A named view puts the camera on that axis looking back at the origin, matching the app's standard views: 'front' is -Y, 'back' is +Y, 'left' is -X, 'right' is +X, 'top' is +Z and 'bottom' is -Z. So 'front' shows the face whose outward normal is -Y. The four isometric views all look down from above, from (+X, -Y, +Z) for 'isometric_front_right', (-X, -Y, +Z) for 'isometric_front_left', (+X, +Y, +Z) for 'isometric_back_right' and (-X, +Y, +Z) for 'isometric_back_left'. Plain 'isometric' is front-right. 2. A dict with "up", "vantage" and "center" keys, each a list of 3 floats in model space: "vantage" is the camera position and "center" the point it looks at. For example {"up": [0, 0, 1], "vantage": [0, -1, 0], "center": [0, 0, 0]} looks at the origin from the front, showing the -Y face. With zoom=True only the direction from "center" to "vantage" matters, because zooming to fit sets the distance: [0, -1, 0] and [0, -200, 0] frame identically. "up" must not be parallel to that direction. 3. 'multiview' for a 2x2 collage of front (top left), right (top right), top (bottom left) and isometric (bottom right).…

NameTypeReqDescription
camera_view–––
highlight_edgesboolean––
max_image_dimensioninteger––
output_path–––
paddingnumber––
session_idstringyes–
zoomboolean––

No output schema declared.

No examples provided.

start_modeling_session ~148

Open an empty CPU modeling session for subsequent tools. Only one modeling session can be open at a time. Stop the current session before starting another. If one is already open, or is still connecting, this fails with an error naming that session's ID so it can be stopped. The server does not expire sessions after an idle or lifetime timeout. Callers are responsible for tracking and enforcing their desired timeout. Pass the returned session_id to execute_kcl to populate the scene, then reuse it with modeling query, selection, and highlight tools. Stop the session explicitly with stop_modeling_session when finished. Returns: str: The session ID to pass to session-aware modeling tools.

Input schema present but exposes no named parameters.

NameTypeReqDescription
resultstringyes–

No examples provided.

stop_modeling_session ~92

Close a persistent modeling websocket session. Also accepts the ID of a session that is still connecting, which cancels that start and frees the slot for a new start_modeling_session call. Args: session_id: The ID returned by start_modeling_session, or the one named by a "already open or starting" error. Returns: None

NameTypeReqDescription
session_idstringyes–
NameTypeReqDescription
resultnullyes–

No examples provided.

submit_job_execution ~58

Start a remote job with an argv command and workspace-relative inputs and outputs. If the response is lost, retry with the same key and uploaded IDs or list existing jobs.

NameTypeReqDescription
idempotency_keystringyes–
submission–yes–
NameTypeReqDescription
completed_at–––
compute_profile_idstringyes–
created_at–––
execution_ended_at–––
idstringyes–
runtime_idstringyes–
started_at–––
state_versioninteger––
statusstringyes–
timeout_secondsintegeryes–

No examples provided.

upload_job_execution_input ~38

Upload a file from the MCP server. Use the returned ID as an input to submit_job_execution.

NameTypeReqDescription
input_filestringyes–
NameTypeReqDescription
byte_lengthintegeryes–
idstringyes–
sha256stringyes–

No examples provided.

use_job_execution_artifact ~87

Read a retained output as paged UTF-8 text or an inline image. To save any format, use download with a new file path on the MCP server. Only text reads accept a byte offset.

NameTypeReqDescription
actionstring––
artifact_idstringyes–
destination–––
job_idstringyes–
offsetinteger––

No output schema declared.

No examples provided.

visualize_sketch ~313

Render a named 2D KCL sketch as a solver-debug PNG. The image shows sketch geometry and solver freedom without requiring an existing modeling session. ``sketch_name`` is the sketch's variable name, such as ``profile`` in ``profile = sketch(on = XY) { ... }``. Use ``get_sketch_constraint_status`` to discover sketch names when needed. A completed sketch may still be rendered after a later execution failure. A returned PNG does not mean the whole project executes successfully. If names repeat (for example, a function called twice), select an ``instance_index`` from a fresh constraint report for the same entrypoint. Args: sketch_name: Variable name of the sketch to render. kcl_code: KCL source code containing the sketch. kcl_path: Path to a KCL file or project containing ``main.kcl``. output_path: If provided, write the PNG to this file or directory and return its absolute path. A directory receives ``image.png``. If omitted, return the PNG inline as ImageContent. instance_index: Zero-based creation order among sketches with this name. Omit for a unique name; refresh the report after edits. Returns: The inline PNG, its saved absolute path, or an error message.

NameTypeReqDescription
instance_index–––
kcl_code–––
kcl_path–––
output_path–––
sketch_namestringyes–

No output schema declared.

No examples provided.

Install

How do I install the io.github.KittyCAD/zoo-mcp server?

io.github.KittyCAD/zoo-mcp runs locally as a PyPI package, launched with uvx zoo_mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

pypi · zoo_mcp

# add to Claude Code
claude mcp add kittycad-zoo-mcp -- uvx zoo_mcp
Common questions

What is the io.github.KittyCAD/zoo-mcp server?

io.github.KittyCAD/zoo-mcp is listed in the public MCP registry as io.github.KittyCAD/zoo-mcp. An MCP server that provides access to the Zoo API for various CAD operations and tools. This page covers its PyPI package (zoo_mcp).

Is the io.github.KittyCAD/zoo-mcp server safe to use?

io.github.KittyCAD/zoo-mcp scores 60 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. It declares no install or post-install scripts. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

What tools does the io.github.KittyCAD/zoo-mcp server expose?

io.github.KittyCAD/zoo-mcp exposes 61 tools: upload_job_execution_input, submit_job_execution, get_job_execution, list_job_executions, read_job_execution_logs, and 56 more. Their descriptions and schemas cost roughly 9,680 tokens of context every time the server is loaded.

Is the io.github.KittyCAD/zoo-mcp server still maintained?

io.github.KittyCAD/zoo-mcp is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

What licence is the io.github.KittyCAD/zoo-mcp server under?

io.github.KittyCAD/zoo-mcp declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

What did VerifyMCP test on the io.github.KittyCAD/zoo-mcp server?

VerifyMCP ran 15 checks on io.github.KittyCAD/zoo-mcp, covering Known CVEs, Install scripts, Dependency health, Source repository and 11 more. One item could not be completed, a gap on our side rather than a finding about io.github.KittyCAD/zoo-mcp: Malware scan. Endpoint Security and Transport & Reachability do not apply to a PyPI package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.