Run arbitrary Python inside Maya (escape hatch).

WARNING: this is NOT a security sandbox. It can read/write files and run system commands. Destructive patterns require allow_dangerous=True.

Evaluate a MEL command inside Maya (escape hatch).

List or toggle tool groups.

Non-core groups (authoring/animation/rendering) start disabled to keep the tool list small; call with action='enable', group='authoring' to reveal them.

Create a polygon primitive (cube/sphere/cylinder/cone/plane/torus).

List scene nodes, optionally filtered by type and name pattern.

Read attributes of a node.

Set one or more attributes on a node.

Delete nodes by name or wildcard pattern.

Parent a node under another, or unparent it to the world.

Duplicate a node, optionally renaming and reparenting the copy.

Rename a node.

Create a new, empty scene.

Open (or reference) a Maya scene file.

Save the current scene (optionally to a new path).

Summarize the current scene: name, path, units, fps, object count, references.

List the currently selected nodes.

Select nodes by name or wildcard pattern.

Check connectivity to a Maya instance and return basic instance info.

Use this to verify the server/plugin link before running other tools.

List connected Maya instances (session_key, versions, scene, tools, active).

Choose which connected Maya instance handles calls that omit session_key.

Useful when several Maya instances are open.

Screenshot the Maya viewport and return the image so you can see the scene.

Use this to visually check modeling progress before/after making changes. The returned path is also included for reuse.

Display a local image file to yourself (e.g. a modeling reference).

Use when the user points you at a reference image: view it first, then model (optionally placing it in the viewport via import_reference_image).