Check whether the Blockbench bridge plugin is connected and inspect its current project.

Generate and apply a polished layered chibi pet (fox, wolf, cat, or rabbit) from a compact art-direction brief. Includes semantic rigging, face details, articulated tail/ears, accessory, textures, locators, and production animation set.

Atomically create or append a precise Blockbench model from groups/bones, cubes, UVs, textures, pixel patches, and animations.

Build a Blockbench model from an image reference already inspected by ChatGPT/another vision-capable MCP client. The caller must translate the attached image into the structured blueprint: palette, semantic bone hierarchy, dimensions, symmetry, compressed geometry primitives, sockets, and animation tracks. Supports boxes, tapered stacks, alternating chains, ragged cloth, crystal clusters, skulls, rib cages, and trimmed armor plates. Use dry_run=true first for complex references; this validates and reports compiled cube counts without changing Blockbench. After build, always capture a four-view turntable and patch discrepancies against the original image.

Atomically refine existing cubes and bones by stable name/UUID after visual review, or remove unwanted cubes. Use this for turntable-driven correction without rebuilding the project.

Create a professional empty rig preset for a weapon, ModelEngine pet, quadruped pet, or humanoid golem, including VFX bones, hitbox, and attachment locators.

Add one advanced animation to the current project, including Molang values, pre/post data points, step/catmullrom/bezier interpolation, and bezier handles.

Read a compact snapshot of the open Blockbench project, including groups, cubes, textures, animations, and bounds.

Audit the active model for invalid cube bounds, missing textures, UV overflow, unrigged cubes, animation loop discontinuity, frame-strip errors, rig conventions, and item display transforms.

Score the active model's art quality using structural lessons from professional weapon, pet, and golem packs: silhouette layering, semantic naming, detail density, symmetry, rig coverage, and animation timing.

List Blockbench model formats and the capabilities/codec of the active format.

Open an uncompressed .bbmodel project from a local path in a new Blockbench project tab for inspection or editing.

Set the active Blockbench preview camera before visual inspection.

Capture the active 3D viewport so the AI can visually verify proportions, silhouette, and colors.

Automatically frame and capture hero, front, side, and back views of the active model for an AI visual refinement loop.

Save the active project as a .bbmodel file on the local machine.

Export using the active format codec, such as a Minecraft Java item JSON, or use codec 'project' for .bbmodel.