Claude Skills for Blender: Reusable Modeling Instructions

Claude cannot open Blender or click its interface, but it can write Python scripts, review modifier setups, suggest step-by-step modeling procedures, and check topology descriptions against standard rules. Packaging those instructions as named skills means you call "Review mesh topology" instead of retyping the same criteria every session.

Tools like fenbs hold the project modeling decisions alongside the skills: which polygon budget applies, which naming convention is in use, which UV workflow the team chose. A connected assistant reads those at the start of a session so the output reflects the project rather than generic Blender advice.

What Claude can help with in Blender work

Claude is most useful for the text-based side of Blender work: writing and reviewing Python scripts for the bpy API, generating step-by-step procedures for a specific modeling task, and checking a text description of a mesh against topology or naming conventions.

It can also help with structure problems. If you describe a node group, a modifier stack, or an armature hierarchy, Claude can ask targeted questions or suggest alternatives based on the goal you describe. That is different from it seeing your scene; you bring the relevant description to the conversation.

Claude cannot view your viewport, render your scene, or run Blender directly. If you need automation that operates inside Blender, a Python script run from the Blender text editor is the right tool. Claude can help you write and review that script.

A skill for writing a repeatable modeling procedure

A modeling procedure skill takes a mesh goal and returns a sequenced list of steps. The skill is most useful for tasks you do repeatedly: hard-surface asset production, retopology, or UVing a specific asset type.

Write a step-by-step Blender modeling procedure for the following asset.

Asset type: {{e.g., low-poly prop, character head, architectural wall section}}
Polygon budget: {{target count or range}}
Pipeline stage: {{e.g., blockout, production mesh, retopology, UV unwrap}}
Software version: {{Blender version}}

Output:
1. Numbered steps in the order a modeler should follow
2. For each step: the Blender tool or technique, any settings to check, and the check condition before moving to the next step
3. Common mistakes to avoid for this asset type
4. Final quality checklist before handoff

Keep steps to what is needed for this pipeline stage only.

A skill for reviewing a Python script for Blender

Blender Python scripts can break between versions, use deprecated API calls, or produce correct-looking output that is slow on large scenes. A review skill checks for those issues before you run the script.

Review the following Blender Python script.

Blender version target: {{version}}
Script purpose: {{what the script is meant to do}}

Script:
{{paste the script}}

Check for:
- Deprecated bpy API calls for this version
- Operations that will be slow on meshes over 50,000 polygons
- Missing error handling (missing object, wrong mode, etc.)
- Hardcoded values that should be parameters
- Anything that modifies the scene without an undo step

Output:
Issue | Severity: High | Medium | Low | Note | Suggested fix
End with: Safe to run: Yes | No | With caution, and one-line reason.

For version-specific API questions, check the Blender Python API documentation↗ (opens in a new tab) directly before relying on Claude's output. Claude's knowledge of specific bpy calls may lag behind the latest release.

Saving project modeling decisions

Every project has choices that the modeling procedure needs to know: the polygon budget, the UV layout conventions, the naming scheme, the target engine and its normal map format. Those choices belong in the project context, not re-explained in each conversation.

Record them once as Decisions and rules in fenbs. When Claude reads the project context at the start of a session, it already knows the polygon budget before you run the procedure skill. The output matches the project constraints from the first call.

Project decisions also include what has been tried and found to not work. If a specific modifier stack caused performance problems on a particular mesh type, recording that as a decision prevents the next session from suggesting the same approach. Negative knowledge, what did not work and why, is as useful as the approved conventions. A modeler returning to a project after three months benefits from both.

Adapting skills for different Blender workflows

Game asset production, architectural visualization, and character modeling each have distinct conventions. A skill for a game prop workflow specifies a polygon budget and LOD stages. A skill for an arch-viz workflow specifies camera-facing polygon distribution and naming for the renderer. A character skill specifies topology flow around deformation areas. A single generic Blender skill satisfies none of these well.

Create a separate skill for each workflow type and name each for its context: "Hard surface game asset procedure", "Arch-viz interior mesh procedure", "Character retopology procedure". A modeler who works across multiple project types calls the skill for the current project rather than adjusting a generic one before each session.

Store each workflow skill in the project it belongs to. If a pipeline change updates the polygon budget for one workflow, that update happens in one project's skill without affecting others. The version history in fenbs shows when the budget changed and provides context for assets produced before that date. A modeler reviewing old assets from a prior pipeline stage can see which constraints were in effect when they were made.

How to do this with fenbs

1. Create a free account and add a project for the production you are working on.
2. Add each skill to Skills and prompts: "Modeling procedure", "Python script review".
3. Add the project constraints as Decisions and rules: polygon budget, naming convention, target engine, UV workflow.
4. Connect Claude over MCP from Integrations. At the start of a session, ask Claude to read the project context before running a skill.
5. When a project decision changes, update the rule in fenbs so every future session reflects the new constraint automatically.

For more skill patterns, see Claude skills examples and how to create a Claude skill. For an overview of how skills fit with Projects and Artifacts, see Claude Projects vs Skills vs Artifacts. The claude skills marketplace and claude skills files cover importing and sharing skills.