scenario-zbrush-expert
Use when an agent drives Maxon ZBrush 2026 for any task (sculpt in ZBrush, DynaMesh, ZRemesher, UVs, export, polypaint, posing, 3D print, clean up an AI mesh in ZBrush), when writing ZBrush Python (zbrush.commands) or ZScript, when the ZBrush bridge is not answering, a palette path or set() does not…
Install / Use
npx skills add scenario-labs/skills --skill scenario-zbrush-expertInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Content & MediaSupported Platforms
Our assessment of scenario-zbrush-expert
scenario-zbrush-expert scores 89/100 on our quality scale, 515th of 1,148 Content & Media skills we index (top 45%).
Its SKILL.md is 21 KB long, well organised into 10 sections with 1 code example: a thorough specification that gives an agent plenty to work with.
With 1,048 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated 35 days ago, so scenario-zbrush-expert is actively maintained.
- Our last check on 2026-11-22 found the source still online.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands.
Automated pattern scan on 2026-11-20. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
scenario-zbrush-expert compared with similar skills
All 4 of these similar skills score higher than scenario-zbrush-expert; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| scenario-zbrush-expert (this skill)by scenario-labs | 89 | 1.0k | 35d ago | SKILL.md |
| Agent-Reachby Panniantong | 100 | 114.0k | today | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 80.5k | today | CLAUDE.md |
| Scraplingby D4Vinci | 100 | 104.9k | 5d ago | MCP Server |
| crawl4aiby unclecode | 100 | 87.7k | 2d ago | MCP Server |
Frequently asked questions
- How do I install scenario-zbrush-expert?
- Run
npx skills add scenario-labs/skills --skill scenario-zbrush-expert. The install tabs above show the steps for each supported agent. - Which AI agents does scenario-zbrush-expert work with?
- It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is scenario-zbrush-expert safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is scenario-zbrush-expert still maintained?
- The repository was last updated 35 days ago, so scenario-zbrush-expert is actively maintained.
Skill content
View source on GitHubname: scenario-zbrush-expert description: "Use when an agent drives Maxon ZBrush 2026 for any task (sculpt in ZBrush, DynaMesh, ZRemesher, UVs, export, polypaint, posing, 3D print, clean up an AI mesh in ZBrush), when writing ZBrush Python (zbrush.commands) or ZScript, when the ZBrush bridge is not answering, a palette path or set() does nothing, a scripted stroke does not sculpt or a dialog blocks a script, or when a ZBrush result has to be judged the way a professional artist would." license: MIT
ZBrush expert (router and agent protocol)
An agent does expert ZBrush work as a loop: deterministic palette operations, synthesized strokes
and canvas renders from fixed views, gated at every stage. One long script is not the method. This
skill is the protocol every ZBrush task follows, the shared toolkit in scripts/, and
the map to the team skills. Target: ZBrush 2026.2.1 on macOS, driven through the project's Python
bridge. If a sibling skill named here is missing from your available skills, load it with
skill_get (Scenario MCP) or ask the user to install it
(https://github.com/scenario-labs/skills#install); unattended, proceed from tool schemas and flag
the gap.
1. Execution channel
| Channel | Use for | Rules |
| --------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Bridge into a running ZBrush (scripts/zb_launch.py, over the client scripts/zb_bridge.py) | all interactive work: palette operations, strokes, renders, measurements | zb_launch.start() spawns ZBrush with ZBRUSH_PLUGIN_PATH=scripts/zb_plugins and ZB_BRIDGE_AUTOSERVE=1. It waits for "plugin loaded", the port and a ping, then quits the ZHomePage helper. Work with zb_launch.run(code) or zb_launch.call("zb_ops", "dynamesh", 256). End with zb_launch.stop(save_to=".../x.ztl"). |
| -script batch launch | unattended jobs on saved files (export every SubTool, reports) | ZBrush.app/Contents/MacOS/ZBrush -script /abs/job.py args /abs/scene.ZPR: the scene is the LAST argument, script args start at sys.argv.index("-script") + 2, report through files and sys.exit(code) (Maxon quickstart). Whether ZBrush quits on exit [verify]. |
Rules for the bridge:
- Main thread only. Every
zbccall runs in main mode, which is the default ofrunandcall. The embedded VM keeps the GIL on ZBrush's main thread, so a socket thread alone starves. The proven design is a serve loop on the main thread that callsupdate(redraw_ui=True)between polls (README). - One driver. Only one agent drives ZBrush at a time. ZBrush is single-instance:
start()refuses a second instance and never kills a ZBrush it did not spawn. - Short, measured calls. Each call returns numbers (faces, volume, file bytes). Long operations such as ZRemesher, Decimation or Divide into millions get minute-scale timeouts. A timeout does not cancel queued code: it still runs when the main thread frees. Ping and read state before sending anything else.
- No blocking UI. Never call
message_*,ask_*orshow_note(..., 0). For visible progress useset_notebar_text(text, progress). - One shared interpreter. ZBrush keeps one embedded Python for the whole session: what a call
leaves behind stays until restart.
runandcallleavesys.path,sys.modules, stdout and the working directory as found and pop every toolkit module, also after an error (Maxon Style Guide; the server runs the restore hook). Never add a folder tosys.pathfor good, never import toolkit code outsidezb_launch.run.zb_launch.hygiene()must return empty lists;purge()removes a leak. - Long sequences. Group deterministic steps into one call:
zb_ops.sequence([["divide", [2]], ["polish", [10]]])runs them insidefreezewithshow_actions(0);quiet()andfrozen(fn)are the parts (SDK GUI reference). Never freeze a plugin press or a canvas export; callupdate(redraw_ui=True)before a snapshot. - Locked screen.
Document:Exportstill writes the canvas PNG (v03);screencaptureshows only the lock screen. At 05:30 on 2026-09-24 a launch stalled after the Home Page press, while the 05:20 launch under the same lock worked.start()detects the stall, terminates the process it spawned and reports a diagnosis (details inzb_launch.py). - Check
zbrush_info(0)= 2026.2 before trusting anything below.
2. Capability matrix (ZBrush 2026.2.1, 2026-09-24)
| Action | Status | Substitute |
| --------------------------------------------------------------------------------- | ---------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------- |
| Press, set and read palette items by path (DynaMesh, Divide, masks, exports) | yes, proven (v01, v03) | always exists() first: set() on a wrong path is silent |
| Synthesized strokes Stroke("(ZObjStrokeV02n<N>=H<x>V<y>...)") + canvas_stroke | yes, proven (v02: volume 4.189 to 4.193) | |
| Replaying a recorded stroke with an offset | yes (v01: volume 4.184 to 4.189); rotation and scale [doc] | |
| canvas_click drags to sculpt | no (v01: no change) | canvas_stroke |
| Dialog-free canvas PNG and OBJ export | yes (v03: 0.04 s, 0.08 s) | |
| ZTL save by path | path exists, dialog-free [verify] | |
| Aiming strokes at model points | front view measured on the v03 OBJ; other views [verify live_04] | zb_stroke.fit_convention |
| Pen pressure in strings | unknown [verify live_02] | Draw:Z Intensity and Draw Size per stroke |
| Smoothing by stroke (Shift) | no: a Smooth brush only becomes the Shift brush | Tool:Deformation:Polish, hPolish brush |
| ZModeler polygon clicks, Gizmo drags, pivot placement | no | Deformation sliders, Geometry X/Y/Z Position and XYZ Size, DynaMesh with Booleans, ZRemesher; else hand to a human or computer-use agent |
| Text prompts (SubTool Rename, New Folder) | no | name parts at creation; set_tool_path renames the top SubTool only |
| FBX export options dialog | no ("You cannot escape any GUI", Maxon forum) | OBJ or .GoZ by path |
| press_key | no ("Does not work" in the stub) | ZScript macro with IKeyPress [verify]; System Events from outside |
| merge_undo, undo control | no ("Not implemented") | versioned ZTL saves, Tool:Morph Target:StoreMT |
| Mask, layer, polygroup data | no API (Maxon forum 2026) | buttons plus visual checks; OBJ export for geometry |
| UV Master and Decimation Master presses | paths exist; return of control [verify live_08, live_09] | native Tool:UV Map:Create (Unwrap):Unwrap |
3. The expert loop in ZBrush
- Brief in numbers. Purpose (film, game, print, concept), polycount per stage, print height in mm, style, references. When the brief is silent, pick the domain skill's documented default and write it down.
- Plan big to small. Block with primitives, ZSpheres or low DynaMesh; then primary, secondary and tertiary forms. Topology, UVs and export come last (Pablo Munoz Gomez's order in Understanding ZBrush's Logic).
- Build one stage with deterministic operations first; they are exact and cheap. Use strokes
only where the form needs them. Before any stroke: Edit mode on, symmetry state chosen on
purpose, brush and Draw values set and read back, view fixed with
set_transform. Record the volume before and after as evidence that the stroke sculpted. - Gate the stage.
- Numbers:
zb_ops.stats(), pluszb_audit.audit()on an OBJ export.
- Numbers:
Truncated for display — read the full file on GitHub.
Related Skills
Agent-Reach
114.0kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
80.5kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Scrapling
104.9k🕷️ The BEST Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ and follow here for daily tips and tricks: https://x.com/Scrapling_dev
crawl4ai
87.7kOfficial Crawl4AI Cloud: a hosted web scraping, extraction, and search API for AI agents. Free API key, 1,000 credits a month, no card. Fully open source: this repo is the Apache-2.0 crawler that powers it, and you can self-host it.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
