agentleFS
zombicozombico/mojulo/.claude/skills/skull-study/SKILL.md

skull-study

Skill4 starsChanged 33 days agoApache-2.0View on GitHub

Render an animal HEAD as a skull-shape turntable — an azimuth sweep (front → ¾ → lateral → back) plus top-down and under views, cropped to the head — so a skull reads correctly from EVERY angle, not just one. Catches the failure modes a single view hides (a hollow/open-tube face that only shows head-on, a pale interior scoop that only shows at ¾, a lopsided or over-long muzzle that only shows from above). Use before/after tuning any animal head, when adding a new species' skull, or after editing the skull/welded-head builders (figure-animal-skull.js, weldedSkull in figure-animal-skin.js). Invoke as `/skull-study <species|archetype>` or `/skull-study --recipe head.json`.

The file

Files
SKILL.md

skull-study

The head is the hardest part of an animal to get right, and the part a single render lies about most. This is the check pass for skull shape: it renders a head from all around so you can Read the turntable and catch what a front-only or lateral-only view hides.

Why a turntable and not one view — learned the hard way on the leopard head:

  • the hollow / open-tube face (a short-muzzled skull with no front cap) reads fine in profile but is a see-through ring head-on;
  • the pale interior scoop (looking into the open muzzle) only appears at ¾;
  • an over-long, lopsided, or wrongly-dropped muzzle only reads from top / under.

Use it

node .claude/skills/skull-study/study.mjs <subject> [<subject>...] [--recipe file.json] [--out dir]

Then Read the printed PNGs — that is the analysis. Each subject writes a turntable to <out>/<subject>/png/ (default /tmp/skull-study):

viewreads
1-frontface plane — hollow ring? eyes/nose seated? countershade chin?
2-front3q · 3-threeqtrthe recognizability angle — muzzle mass, cheek fill, scoop
4-lateralprofile — muzzle length, brow/stop, nasal bridge, jaw
5-rear3q · 6-backcranium closure, ear seating
7-topdorsal skull, muzzle centering + symmetry, nasal bridge prominence
8-undermuzzle / jaw underside, chin

Subjects:

  • a species — any ZOO_BUILDS key (wolf, cougar, lion, deer, fox, …).
  • an archetype — any QUADRUPED_ARCHETYPES key (feline, canine, stumpy, equine, rodent, …), rendered with neutral skin+coat so you see the RAW skull.
  • --recipe file.json — an ad-hoc { archetype, opts } head being TUNED before it's graduated into ZOO_BUILDS. The file's basename names the subject. Repeatable.

Examples:

  • … study.mjs cougar — the existing puma head, all around.
  • … study.mjs feline canine stumpy — compare three raw archetype skulls.
  • … study.mjs --recipe /path/leopard-recipe.json — a head under construction.

The skull dials it's checking (map)

Reference for what moves the skull, so a turntable finding maps to a dial. Skull cfg (skullCfg), consumed by protoSkull (overlap) and weldedSkull (skin):

  • length overall skull length · width overall breadth · dome cranium height
  • muzzle muzzle fraction of length (↑ = longer snout) · snout snout-tip radius (↑ = blunter/wider tip) · muzzleDrop how far the nose dips · jaw lower-jaw size
  • boxy superellipse squaring (protoSkull only; the welded head is round)

Welded-head-only muzzle dials (opt-in; default neutral, so long-muzzled animals are untouched) — added for brachycephalic/cat faces:

  • pad whisker-pad masses that WIDEN the front into a squared muzzle (small: the cap radius tracks the opening, so a big pad blows the whole muzzle up).
  • bridge a raised nasal-bridge ridge up the top-centre, between the eyes.
  • capDepth front-cap forward depth (<1 = flatter, SHORTER, squarer front).

Face decorators (face cfg, faceDecor): eyeR/eyeSide/eyeUp/eyeSink, noseR/noseFwd, earLen/earW/earTip/earUp/earSide.

Invariant learned: the welded head is a marchAxis lathe tube — CLOSE its open front AT THE MARCH (the weldedSkull front cap), never by overlaying a second surface (overlays z-fight and reveal the interior). See animal-from-dream.plan.md gap ledger.

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.