agentleFS
witnesstodarkwitnesstodark/mr-mak-workspace/.claude/skills/motion-reference-workflow/SKILL.md

motion-reference-workflow

Skill718 starsChanged 22 days agoMITView on GitHub

Prepare and select character motion references, or transfer approved motion to a human reference for a later mocap workflow.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the approved character identity.

    line 13
  2. Distinctive:

    Keep a small take ledger: source, prompt or edit intent, provider/model, local result, observed problems and decision.

    line 33

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Motion references

Resolve whether the user needs a character reference video for authored animation or human footage for a later mocap tool. The camera follows that purpose. Character-first and human-first are both valid; preserve the user's chosen route.

Establish the opening pose, action, contact point, recovery, duration and camera. Use the approved character identity. Before a new video, verify its first-frame pose instead of relying on a model to invent the correct start. Keep framing stable enough to read the relevant limbs and contacts.

Choose the provider/model from the user's selection or current discovery. Use the fal or Higgsfield module for execution and recovery. The number of candidates comes from the task; do not impose a costly fixed batch.

Review support changes, strike readability, foot contact, impossible joints, interpenetration, unwanted camera motion and loop seams. Keep candidate names and decisions stable. Preserve longer source motion when creating a shorter gameplay variant, and record event times in seconds as well as frames.

For an appearance transfer, use the accepted video as motion authority and the approved human/reference image for appearance. Preserve camera and timing; do not recreate the action from text. Check the transferred limbs and contacts. Generated human-looking footage is not itself captured skeleton data. A later mocap extraction and its quality checks remain separate work.

Keep a small take ledger: source, prompt or edit intent, provider/model, local result, observed problems and decision. Distinguish generated, reviewed, selected and approved. Carry the selected take's identity into the animation handoff. For a creature-to-human transfer, verify which limb represents each contact and reject anatomy changes that destroy the original action. Preserve the original and any timed trim; do not replace a motion reference with a new performance merely because its appearance is cleaner.

Example: a two-second robot walk loop with fixed camera and readable foot contacts. A successful clip must loop cleanly and retain the approved robot; remote completion alone does not establish those properties.

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.