Claude Cursor Skill

meshy-3d-generation

Generates and edits digital 3D assets with the Meshy CLI: text-to-3D and image-to-3D models, texturing, remeshing and LODs, UV unwrapping, format conversion (GLB, FBX, OBJ, USDZ, STL, 3MF, Blender), rigging, animation, motion clips from text and reference images. Use when the use

LLM Mart · 0 points · 3 views 166 listing impressions 0 install-command copies
Download meshy-dev-meshy-3d-agent-skills_meshy-3d-generation-7f8e331.zip · 27 KB

Install

skills CLI npx skills add https://github.com/meshy-dev/meshy-3d-agent/tree/main/skills/meshy-3d-generation
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install meshy-dev-meshy-3d-agent@llmmart
Git git clone https://github.com/meshy-dev/meshy-3d-agent.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole meshy-dev/meshy-3d-agent collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Meshy 3D Generation

Use Meshy CLI 0.8 (0.8.6 or later) to turn the user's request into a delivered asset. Completion means saving the requested files in the requested location and inspecting an available preview. Login or task submission alone is not completion. Choose the shortest suitable pipeline and reuse existing assets for follow-ups.

Read what this request needs

Need Reference
CLI runner, credentials, or output location Setup
Generation and the models to pair for a project, texturing, remeshing, conversion, scale, or 2D images Pipelines
Rigging, animation, or generated motion Animation
Cost, waiting, preview, delivery, or finding a previous model Delivery
A chat that shows images or web pages (desktop or IDE agent apps), at delivery or when the user asks to see the model Visual hosts
Failed or uncertain operation Troubleshooting

Read the relevant section when needed; recipes are examples to adapt to the requested result, not a requirement to execute every stage. An untextured model skips texture generation; an existing GLB needs no extra format conversion. Use command help when parameters are uncertain.

First use and authorization

In DeepSeek Harness with the Meshy plugin (DSH_MESHY_CLI is set) run every command as "$DSH_MESHY_NODE" "$DSH_MESHY_CLI" …, never a meshy on PATH, install nothing, and sign in with the meshy_sign_in tool. Everywhere else run meshy --version before the first command, without asking; when it is missing or older than 0.8.6, install it with npm install -g meshy-cli@0.8 and say so in one line (Setup, section 1).

Reuse a verified session. If login is needed, run auth login --device (with --region cn for a user in mainland China) as Setup, section 3 describes: show its URL and user code in commentary/progress, not a final answer, wait on that same command in this turn, verify auth status, then continue the original request. Browser approval stays with the user; credentials never belong in chat, and only the CLI touches its config directory.

Execution boundaries

  • Run API, upload, download and mesh operations through the CLI. Local inputs have a 50 MiB limit, and a local model must be GLB or glTF (rigging, UV unwrap and multicolor: GLB); use the supported URL input route for anything else.
  • Use --format json --no-update-check on every command: stdout is one meshy.cli/v1 envelope, auth included. Resolve WORKSPACE from the user's output path and pass it to writes.
  • Know what the model is for before the first paid create: a game asset, a character to rig, a render, a 3D print. If the user hasn't said, ask once; the answer picks the route and the models (Pipelines). A game asset or a face-count budget means smart topology, not a standard model remeshed down.
  • Submit once with create --async, retain the owning resource and ID, then wait on that task. Keep the agent turn active while awaiting a running command. Follow Troubleshooting on timeout or unknown submission; these do not authorize another paid create.
  • Without a usable photo, recommend making a reference image first and building the model from it; text-to-3d is the quicker alternative (Pipelines, Start from an image).
  • Continue work within the user's approved scope and budget. Propose additional paid stages or reruns before spending; Delivery specifies estimates and balance checks.

Deliver

Download only needed assets. Inspect and show a preview when available; a missing preview must not prevent delivery or be described as inspected. Delivery covers earlier-stage preview lookup and actual charges. Report the result and any concrete limitation; a rendered image does not establish topology quality or provide a 3D viewer. Where your chat shows images or web pages, Visual hosts covers showing the preview and the model there.

Files (meshy-3d-agent)
  • references
    • animation.md 4.3 KB
      Loading…
    • delivery.md 9.2 KB
      Loading…
    • pipelines.md 20.6 KB
      Loading…
    • setup.md 13.9 KB
      Loading…
    • troubleshooting.md 9.6 KB
      Loading…
    • visual-hosts.md 1.7 KB
      Loading…
  • SKILL.md 4.7 KB
    ---
    name: meshy-3d-generation
    description: "Generates and edits digital 3D assets with the Meshy CLI: text-to-3D and image-to-3D models, texturing, remeshing and LODs, UV unwrapping, format conversion (GLB, FBX, OBJ, USDZ, STL, 3MF, Blender), rigging, animation, motion clips from text and reference images. Use when the user asks to make, texture, rig, animate or convert a 3D model or game asset. For physical 3D printing use meshy-3d-printing."
    license: MIT
    compatibility: "Needs Node.js 22.12+ and the Meshy CLI 0.8.x, 0.8.6 or later (npm install -g meshy-cli@0.8); network access to the Meshy API."
    ---
    
    # Meshy 3D Generation
    
    Use Meshy CLI **0.8** (0.8.6 or later) to turn the user's request into a delivered asset. Completion means
    saving the requested files in the requested location and inspecting an available preview.
    Login or task submission alone is not completion. Choose the shortest suitable pipeline and
    reuse existing assets for follow-ups.
    
    ## Read what this request needs
    
    | Need | Reference |
    |---|---|
    | CLI runner, credentials, or output location | [Setup](references/setup.md) |
    | Generation and the models to pair for a project, texturing, remeshing, conversion, scale, or 2D images | [Pipelines](references/pipelines.md) |
    | Rigging, animation, or generated motion | [Animation](references/animation.md) |
    | Cost, waiting, preview, delivery, or finding a previous model | [Delivery](references/delivery.md) |
    | A chat that shows images or web pages (desktop or IDE agent apps), at delivery or when the user asks to see the model | [Visual hosts](references/visual-hosts.md) |
    | Failed or uncertain operation | [Troubleshooting](references/troubleshooting.md) |
    
    Read the relevant section when needed; recipes are examples to adapt to the requested result,
    not a requirement to execute every stage. An untextured model skips texture generation; an
    existing GLB needs no extra format conversion. Use command help when parameters are uncertain.
    
    ## First use and authorization
    
    In DeepSeek Harness with the Meshy plugin (`DSH_MESHY_CLI` is set) run every command as
    `"$DSH_MESHY_NODE" "$DSH_MESHY_CLI" …`, never a `meshy` on PATH, install nothing, and sign in with
    the `meshy_sign_in` tool. Everywhere else run `meshy --version` before the first command, without
    asking; when it is missing or older than 0.8.6, install it with `npm install -g meshy-cli@0.8`
    and say so in one line ([Setup](references/setup.md), section 1).
    
    Reuse a verified session. If login is needed, run `auth login --device` (with `--region cn` for a
    user in mainland China) as Setup, section 3 describes: show its URL and user code in
    **commentary/progress, not a final answer**, wait on that same command in this turn, verify
    `auth status`, then continue the original request. Browser approval stays with the user;
    credentials never belong in chat, and only the CLI touches its config directory.
    
    ## Execution boundaries
    
    - Run API, upload, download and mesh operations through the CLI. Local inputs have a 50 MiB
      limit, and a local model must be GLB or glTF (rigging, UV unwrap and multicolor: GLB); use the supported
      URL input route for anything else.
    - Use `--format json --no-update-check` on every command: stdout is one `meshy.cli/v1` envelope,
      auth included. Resolve `WORKSPACE` from the user's output
      path and pass it to writes.
    - Know what the model is for before the first paid create: a game asset, a character to rig, a
      render, a 3D print. If the user hasn't said, ask once; the answer picks the route and the models
      ([Pipelines](references/pipelines.md)). A game asset or a face-count budget means smart topology,
      not a standard model remeshed down.
    - Submit once with `create --async`, retain the owning resource and ID, then `wait` on that
      task. Keep the agent turn active while awaiting a running command. Follow Troubleshooting
      on timeout or unknown submission; these do not authorize another paid create.
    - Without a usable photo, recommend making a reference image first and building the model
      from it; text-to-3d is the quicker alternative ([Pipelines](references/pipelines.md), *Start
      from an image*).
    - Continue work within the user's approved scope and budget. Propose additional paid stages
      or reruns before spending; Delivery specifies estimates and balance checks.
    
    ## Deliver
    
    Download only needed assets. Inspect and show a preview when available; a missing preview
    must not prevent delivery or be described as inspected. Delivery covers earlier-stage preview
    lookup and actual charges. Report the result and any concrete
    limitation; a rendered image does not establish topology quality or provide a 3D viewer.
    Where your chat shows images or web pages, [Visual hosts](references/visual-hosts.md) covers
    showing the preview and the model there.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.