agentleFS
JiamingShi-JsPhysJiamingShi-JsPhys/3d-fabric/skills/reference-look-calibration/SKILL.md

reference-look-calibration

Skill2 starsChanged 3 days agoMITView on GitHub

Calibrate Blender materials, lighting, camera crop, emission/glow, color management, and aura/HUD styling against supplied original/reference images using measurable color, brightness, saturation, bbox, and mask statistics. Use when a product must match the lighting/look of a source image or when audits say the product is too bright, desaturated, wrong hue, wrong glow, or wrong silhouette extent.

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 5 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use when: Reference-based look-dev, lighting/material/color calibration, original-image visual matching, glow/aura/HUD matching, …

    line 10
  2. Distinctive:

    Before the first Blender call, read ../text-to-blender/references/mcp-backends.md.

    line 12
  3. Very specific:

    Use scripts/look_fit_report.py to compare:

    line 26
  4. Distinctive:

    Do not tune lighting before geometry/UV fit gates pass.

    line 47
  5. Distinctive:

    Do not change accent/glow hue away from the source reference unless the user explicitly asks.

    line 48
  6. Distinctive:

    Do not hide mismatched geometry with bloom, fog, or overexposure.

    line 49
  7. Distinctive:

    Write look metrics before and after calibration.

    line 50

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

MCP backend

Use when: Reference-based look-dev, lighting/material/color calibration, original-image visual matching, glow/aura/HUD matching, HSV/brightness/mask fit reports, final look pass after geometry and UV are locked.

Before the first Blender call, read ../text-to-blender/references/mcp-backends.md. Prefer official MCP for shared capabilities; preserve the selected backend for the whole task. Tool names in recipes are client/legacy examples: map them through that reference and its agent-host guide.

Reference Look Calibration

This skill handles the final visual match after geometry and UVs are locked. It must not compensate for wrong shape, wrong UVs, or wrong camera framing.

Inputs

  • accepted geometry / UV fit reports;
  • reference/original image(s);
  • product preview render(s);
  • optional mask overlays and multiview fit reports.

Metrics

Use scripts/look_fit_report.py to compare:

  • object/bright-mask IoU and bbox extent;
  • configured hue-band mask IoU for glow/accent assets;
  • mean/std HSV and BGR under object mask;
  • brightness p95 / mean;
  • saturation mean;
  • hue drift.

Calibration order

  1. Camera crop / orthographic scale — object extent must match before color tuning.
  2. World/background darkness.
  3. Base material hue/value/saturation.
  4. Emission strength and glow color.
  5. Aura/HUD ring color and thickness.
  6. Color management / exposure / gamma.
  7. Final render overlay and metrics.

Hard rules

  • Do not tune lighting before geometry/UV fit gates pass.
  • Do not change accent/glow hue away from the source reference unless the user explicitly asks.
  • Do not hide mismatched geometry with bloom, fog, or overexposure.
  • Write look metrics before and after calibration.

Handoff

Consumes validation/multiview_fit_report*.json and product previews. Produces validation/look_fit_report.json, look overlays, and updated material/light/render settings.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.