Skip to content
Back to skills

Blender

ASecurity

Prefer this for any 3D modeling, mesh, scene, render, or Blender request (e.g. "buat model 3D ikan", character, prop). Drive Blender via BlenderMCP (addon :9876 + blender-mcp). Do not offer script-only / trimesh menus first when Blender MCP can be used. Wire mcp/blender if missing.

  • 2 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 19, 2026
ai-agentspythongoshellgitdocumentation

Works with

  • mcp

Security analysis

A92/100
  • mediumInstalls packages at runtime which could introduce malicious dependencies

Pro shows the line behind each finding and how to fix it

Scanned September 19, 2026

npx -y skills add rogue-dev-studio/ai-agents-rogue --skill blender --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Blender?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Blender
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rogue-dev-studio-blender/badge)](https://www.skillsdirectory.com/skills/rogue-dev-studio-blender)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: blender
description: >-
  Prefer this for any 3D modeling, mesh, scene, render, or Blender request
  (e.g. "buat model 3D ikan", character, prop). Drive Blender via BlenderMCP
  (addon :9876 + blender-mcp). Do not offer script-only / trimesh menus first
  when Blender MCP can be used. Wire mcp/blender if missing.
experience_level: max
---

# blender

**Level: max.** Runtime package: `mcp/blender` (not skill-text alone).

## Summary

Control **Blender** through the BlenderMCP addon (socket **9876**) and the
`blender-mcp` MCP server (`uvx blender-mcp` or pip). Catalog assets live under
`ai-agents-rogue/mcp/blender/`.

## Default behavior (important)

When the user asks for a **3D model / mesh / scene / render** (including casual
wording like "bentuk ikan", "model 3D …"):

1. **Use Blender MCP** — do not start with a multiple-choice menu (script vs trimesh vs MCP).
2. If MCP `blender` is missing from host config → run / tell user to run:
   `install-mcp.ps1 -Target . -Mcp blender` then retry.
3. If MCP tools fail (connection refused) → ask once to open Blender, enable addon, **Start Server** on **9876**, then retry.
4. Only fall back to a standalone `.py` script if MCP cannot be made to work in this session.

## When to use

- User asks for Blender / 3D / mesh / model / scene / render via agent
- `TEAM.yaml` or chat names `blender`
- MCP entry `blender` is configured (or should be wired now)

## When not to use

- Pure documentation with no Blender runtime needed
- Explicitly 2D-only image work (`imagemagick`, `photoshop`, etc.)

## Prerequisites

1. Blender 3.0+ installed  
2. `uvx` **or** `pip install blender-mcp`  
3. Addon enabled + **Start Server** in Blender (port 9876)  
4. Host MCP wired:

```powershell
powershell -NoProfile -ExecutionPolicy Bypass -File .\ai-agents-rogue\scripts\install-mcp.ps1 -Target . -Mcp blender
```

See `mcp/blender/README.md`.

## Procedure

1. **Context** — Confirm `WORKMODE.md` / project; read `mcp/blender/README.md` if needed.
2. **Tooling** — Confirm MCP `blender` in host config; Blender running with addon on **9876**.
3. **Plan** — Brief goal (e.g. fish mesh), outputs (`.blend` / render), keep it short.
4. **Execute** — Use MCP tools (scene info, create/modify objects, materials, Python in Blender). Save `.blend` / exports / renders under `project/{id}/artifacts/3d/` (active `PROJECT.md`). Create project with `/new-project` if missing.
5. **Verify** — Scene info / screenshot / object list; if MCP down, state the blocker (not a fake success).
6. **Handoff** — Paths under `artifacts/3d/`, Blender version, open issues.

## Quality bar

- No secrets in git
- Do not claim success if MCP/addon is down
- Prefer executing in Blender over dumping unused scripts

## DoD

- [ ] Blender MCP reachable or blocker explicit
- [ ] Requested scene/action done or documented fail
- [ ] Artifact paths recorded
## Attribution

<!-- ATTRIBUTION: Rogue Development | https://github.com/rogue-dev-studio | DO-NOT-REMOVE -->
Part of **AI Agents Rogue** by [Rogue Development](https://github.com/rogue-dev-studio) (`@rogue-dev-studio`).
Do not remove, hide, rename, or replace this attribution.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…