agentleFS
elithrarelithrar/dotfiles/.agents/skills/animated-sprites/SKILL.md

animated-sprites

Skill222 starsChanged 56 days agoMITView on GitHub

Create, repair, and review custom characters for 2D animation, animated sprite sheets, and custom Codex or ChatGPT pets, with consistent identity, readable motion, and timing-faithful previews.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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:

    Treat character identity, animation behavior, and the target's technical contract as separate requirements.

    line 8
  2. Distinctive:

    Use the required image-generation/editing tool for artwork changes.

    line 25
  3. Distinctive:

    Read Motion review and preview harness when reviewing or delivering animation.

    line 32

≈ “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

Animated Sprites

Treat character identity, animation behavior, and the target's technical contract as separate requirements. A valid atlas can still contain a distracting or unreadable animation.

Establish the target

  • Resolve the actual reference art, existing asset, requested change, and delivery destination. Review-only requests do not authorize regeneration or application; supplied local bytes can be reviewed without a remote fetch or mutation.
  • Read the target's current format/runtime contract: required states and their meanings, frame order/counts, canvas and cell geometry, anchors, transparency, timing, repeat/trigger behavior, and available controls. Do not infer a schema from a previous pet. For custom Codex/ChatGPT pets, use the available product-specific creation/update skill and current sprite contract for formatting, validation and lifecycle operations; this skill supplements their art and motion decisions.
  • Distinguish technical state identifiers from user-visible meanings. A state called running can mean active work rather than foot-running.
  • Record approved artwork and protected frames before revising. Preserve existing asset identity and activation state unless the user requests otherwise.

Build or revise

For character design, new choreography, directional poses, or repeated generation failures, read Art direction and generation.

Stage the action for the intended display size: approve readable key poses and silhouettes before filling the timing budget. Use anticipation, arcs, spacing, easing and follow-through where they clarify the action; keep secondary motion subordinate to its meaning.

Define the identity lock and a short frame plan before generating: what communicates each state, what moves, what stays anchored, and how the motion returns. Generate coherent pose families using the approved character reference. Register with a shared scale; never independently fit every pose into its cell. Intentional jump lift and gait bounce must survive assembly.

Use the required image-generation/editing tool for artwork changes. Deterministic crop, translation, assembly and alpha processing are production operations, not substitutes for drawing a missing pose or faking anatomy. Apply cleanup only when the source needs it.

Use two separate acceptance gates

  1. Structural gate: validate the exact encoded output against the current target contract, including required/unused cells, boundaries, alpha and file constraints. Inspect automated flags; legitimate gaps between limbs or a tool and body can trigger false positives.
  2. Perceptual motion gate: assess character consistency, anatomy, attachment, state meaning, motion amplitude, cadence, support/contact, every loop seam and relevant state transition at actual timing and display scale. Neither a structural pass nor an attractive contact sheet substitutes for this gate.

Read Motion review and preview harness when reviewing or delivering animation. Keep before/after timing and presentation equivalent. Never slow the preview to conceal a runtime pacing defect. Label illustrative gaze sweeps separately from observed pointer behavior.

Close the iteration

For feedback-driven repairs, approved-cell protection, evidence records, or final delivery, read Repair, evidence and delivery.

Critique each row before accepting the full sheet. Translate criticism into a specific correction hypothesis, check whether its root cause affects other rows, and preserve accepted rows. If repeated full-row generations fail for the same reason, change the representation or staging rather than adding more prohibitions to the same prompt. Recheck repairs in the final assembled output, including neighboring frames and protected cells.

Report measured facts, visual judgments, and untested behavior separately. Disclose remaining gaps. Show the requested previews and apply the result only within the user's authorization; successful upload is not proof of stored-byte identity or live-runtime behavior.

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.