ai-skills-blender
AGENTS.md11 starsChanged 36 days agoMITView on GitHub
When adding or changing any example .blend, maintain a matching self-contained GLB alongside it and update documentation/models/index.json. Follow documentation/models/README.md and the export/catalog scripts in scripts/. Retain all source scenes, including historical variants. Embed required textures, preserve applicable animation, and bake procedural surfaces when glTF cannot represent them directly. Validate the GLB structure and inspect it in the browser; file existence is not sufficient. Record export limitations in .export.json and catalog warnings. Metadata must cite the actual source manifest/review/prompt; do not…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 33,315 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository30 other files it gives its agents.
- blender-convert-2d-3dSkill
- blender-convert-3d-2dSkill
- blender-create-environmentSkill
- blender-create-greyboxSkill
- blender-create-modelSkill
- blender-create-visual-target-2dSkill
- blender-game-exportSkill
- blender-light-cameraSkill
- blender-materialsSkill
- blender-procedural-geometrySkill
- blender-renderSkill
- blender-review-optimizeSkill
- blender-rig-animateSkill
- blender-setupSkill
- blender-uv-bakeSkill
- blender-convert-2d-3dSkill
- blender-convert-3d-2dSkill
- blender-create-environmentSkill
- blender-create-greyboxSkill
- blender-create-modelSkill
- blender-create-visual-target-2dSkill
- blender-game-exportSkill
- blender-light-cameraSkill
- blender-materialsSkill
- blender-procedural-geometrySkill
- blender-renderSkill
- blender-review-optimizeSkill
- blender-rig-animateSkill
- blender-setupSkill
- blender-uv-bakeSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 5
When adding or changing any example .blend, maintain a matching self-contained GLB alongside it and update documentation/models/index.json.
≈ “Common” is an estimate: at least 1 in 5 of the 42,283 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Windows execution priority
Browser model deliverables
When adding or changing any example .blend, maintain a matching self-contained
GLB alongside it and update documentation/models/index.json. Follow
documentation/models/README.md and the export/catalog scripts in scripts/.
Retain all source scenes, including historical variants. Embed required textures,
preserve applicable animation, and bake procedural surfaces when glTF cannot
represent them directly. Validate the GLB structure and inspect it in the browser;
file existence is not sufficient. Record export limitations in .export.json
and catalog warnings. Metadata must cite the actual source manifest/review/prompt;
do not invent missing fields or confuse AI visual-target metadata with model data.
Regenerate exports when source hashes change, commit the catalog with its assets,
and verify anonymous public HTTP access after publication. The independent Model
Viewer consumes these files remotely and must never bundle them in its releases.
Process execution
- Create no auxiliary command windows, including transient flashes. Prefer the existing MCP connection. Before any helper launch, establish no-console creation for the outer launcher and every descendant. If a required boundary is unknown or unsupported, do not launch it: report the operation as blocked and continue independent safe work. Never retry without suppression or open a terminal to work around a failure.
- Keep existing Blender skills, setup adapters, helper scripts, and generated client rules consistent with this priority. Preserve captured diagnostics, bounded execution, and cleanup of owned processes. An absolute executable path,
cmd /c, Blender--background, or hidden-window styling alone does not establish no-console execution. Do not hide, restore, focus, or close the user's existing Blender editor to satisfy this rule.
Blender editor capture preference
Editor screenshots are authorized as part of requested Blender work. Check that Blender is running with a visible, non-minimized editor before capturing; maximization is not required. Both official full-window and area captures have returned black while minimized on this workstation, so minimized capture is not a verified capability. Report "Restore Blender from the taskbar" when this prerequisite fails. Read-only setup checks must not restore or focus windows themselves. Always inspect captured pixels; report capture failures honestly and never substitute a clean render or old screenshot for fresh editor evidence.
Before every Blender editor/window/area screenshot call, freshly check that the relevant Blender process has a visible, non-minimized editor window; do not reuse an earlier setup result. If Blender is closed, prompt the user to open it. If minimized or hidden, prompt the user to restore/show it (maximization is unnecessary). If inspection is unavailable, report the unknown state and ask the user to make the editor visible. Pause screenshot attempts until the issue is resolved, then repeat the window-state check before capturing. Do not automatically restore/focus the window. This prerequisite applies to editor screenshots, not saved render output. A passing check still requires inspection of the captured pixels; if the image is black or stale, report the failure and prompt the user to restore/show the editor before a fresh check and retry.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

