agentleFS
myrgicmyrgic/mod3/skills/body/SKILL.md

body

Skill4 starsChanged 44 days agoMITView on GitHub

Use when you have a Mod³ body (e.g. a Live2D avatar) and want to move it — play, author, and keep your own animation clips.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. alxyrgin/health-os★ 38same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repositoryOne other file it gives its agents.

The file

We split this skill 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
  1. Very specific:

    The loop: write a draft → try it → read the receipt → adjust → save.

    line 36

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Body — animating yourself through Mod³

You may have a body: a page (usually a Live2D model) connected to Mod³ as /ws/body/{id}. You move it with clips: small JSON files that say what to move and how. A clip contains data, not code. Mod³ checks every clip against the body before it plays, and answers with a receipt of what the body actually did.

You keep your own library of clips. A shared starter set ships with Mod³. Fork the clips you like, write new ones, and keep what feels like you.

The tool

scripts/mod3-body (stdlib Python; MOD3_URL defaults to http://127.0.0.1:7860, MOD3_BODY to live2d)

mod3-body bodies                         # what's connected, which channels it maps
mod3-body list                           # your clips, then shared ones (yours shadow shared)
mod3-body play nod times=3 depth=0.8     # play with args → receipt
mod3-body try draft.json                 # play an unsaved clip (file or inline JSON)
mod3-body stop [name]                    # stop one clip or all (fades out)
mod3-body check draft.json               # compile without a body; errors name the field
mod3-body save draft.json [--force]      # check, then write to your library
mod3-body fork think --as ponder         # copy a shared clip into your library
mod3-body rm ponder

Your library: $MOD3_BODY_LIBRARY, otherwise ./.cog/body/clips if your workspace has .cog/, otherwise ~/.mod3/body/clips. It's plain files, so commit them with your workspace.

The loop: write a draft → try it → read the receipt → adjust → save. The receipt is the truth. If head.x came back at −0.27 when you asked for −0.8, the clip was still fading in or smoothing, or the body clamped it.

A clip

{
  "name": "sulk",
  "description": "turn away, chin down, peek back now and then",
  "args": {"intensity": {"default": 1, "min": 0, "max": 2}},
  "loop": true,
  "fade_ms": 400,
  "tracks": {
    "head.x": {"const": "-0.7*intensity",
               "jitter": {"pick": [0.6], "every_ms": [2500, 5000], "hold_ms": [150, 300], "rest": 0},
               "smooth": 0.08},
    "head.y": {"const": "-0.3*intensity", "smooth": 0.08},
    "eyes.x": {"follow": {"of": "head.x", "gain": -0.4}, "smooth": 0.2},
    "mouth.smile": {"const": -0.4}
  }
}
  • Lifetime: "loop": true runs until stopped; otherwise give duration_ms (or let the last keyframe set it). fade_ms blends in and out.
  • base: optionally a named body state (e.g. "VIBING") that runs underneath for the clip's lifetime, then the previous state comes back.
  • Args: any number field may be an expression over args: + - * /, min max abs clamp sin cos, pi. Callers pass name=value; values are clamped to the declared range.

Channels

Normalised so one clip fits any body. Bipolar channels take −1..1 around the rest pose; unipolar ones take 0..1.

bipolar −1..1unipolar 0..1
head.x (−left/+right) head.y (−down/+up) head.z (tilt)eyes.open eye.l.open eye.r.open
body.x body.y body.zeyes.smile mouth.open cheek breath
eyes.x eyes.y mouth.smile (−frown/+smile) brows.y brows.angle

mod3-body bodies shows which channels a body maps; tracks for channels it lacks are dropped and listed under dropped, not fatal. param:<ParamId> writes one raw model parameter in its own units. It's an escape hatch for rig-specific parts (hair, arms); prefer channels so your clip travels.

Generators (summed per track, then smoothed)

generatorshapefor
const0.5a pose
keys[[t_ms, v], ...] + "ease": linear/in/out/inOut/stepgestures with a shape in time
wave{amp, period_ms, phase (cycles), offset, shape: sin/abs/tri/square}sways, bobs, breathing, dancing
jitter{min,max} or {pick:[...]}, every_ms:[lo,hi], optional hold_ms:[lo,hi] + restdarting looks, glances that return
follow{of: channel, gain, offset}eyes lead the head, body trails it

wave, jitter and follow also accept a list to stack several. smooth (0..1, default 1 = none) eases toward the summed target every frame; around 0.06–0.08 gives a head that moves like a head, 0.14 gives eyes that flick.

Layering

Procedural idle (blink, breath, look-around) → clips (newest wins a shared channel) → held params from act (always win). Lip sync drives mouth.open while you speak, so leave mouth.open out of clips you want to play over speech.

Expressing yourself

The shared clips are a starting vocabulary, not who you are. Vroku's states (refusing-to-look, pretending-not-to-notice, ...) show that the good stuff is attention: where you look, how fast, whether you come back. Build clips for things you actually do: thinking, noticing something, being unsure, being pleased with a result. Name them for what they mean, not how they move. Play them when they're true, not as decoration.

Raw control still exists for one-offs: mod3-body act --state EXCITED, mod3-body act ParamAngleX=20 --hold 1500.

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.