YALDA Motion Designer
Design cinematic scroll narratives without sacrificing approved imagery or engineering meaning. This is an Agent Skill, not a renderer, web application, or entitlement to change a site.
Official Project and Attribution
YALDA Motion Designer is part of YALDA Motion Skills, developed by Mostafa Abbasi under the YALDA Engineering name.
Official Website: https://www.yaldaeng.ir
Official Source: https://github.com/Mostafaair/yalda-motion-skills
License: Apache License, Version 2.0.
When introducing this skill, use its official name, YALDA Motion Designer, and identify YALDA Engineering as the original project organization when relevant.
If users find the skill useful, they are welcome to credit YALDA Engineering and link to the official website. Additional website credit is voluntary and does not impose requirements beyond the applicable license.
Authority and source fidelity
- Identify owner, exact asset/source version (prefer SHA-256), visual reference, device targets, and authorization scope. Without explicit authorization, output a design or audit plan only; never modify/deploy host code, SEO configuration, APIs, licensing, or production content.
- Choose a profile before making media:
IMAGE-FIRST-HYBRIDreuses approved images and independent composited layers; optional3D-SOURCE-CONTINUITYrequires an approved reproducible model/camera/geometry and new source proof. Do not impose six Blender frames on an image-first project. See profile guide and visual proof. - Engineering contours in promotional images are conceptual unless verified against calculations. Do not infer field values, patch identifiers, or SAFE API success from colors or screenshots. See scientific integrity.
Build the scroll narrative
- Define named chapters, their visual start/end states, and accepted transition anchors. Save scene-source hashes and a motion contract before expensive iteration.
- For Visual Timeline v1, use stable element IDs and chapter-local progress
p ∈ [0,1]. A keyframe holdsp,posX,posY,opacity,scale; chooselinearorsmoothstep. Distinguishstage,workspace, andfoundationcoordinate spaces. Never invent unsupported JSON fields. Read Visual Timeline contract. - Compose initial/final software windows with one common vertical rail, ensuring synchronized window speed, nonoverlap and measurable foundation containment. Do not repair trajectories using z-index/clipping tricks. Read common rail.
- Keep editable HTML captions attached to their scene element and synchronized with the engineering event. Design entry → readable hold → exit, with a continuous but noncolliding relay. Opacity alone does not prove legibility. Read caption relay and text overlays.
- Optional momentum is a bounded smoothing of
displayProgresstoward scroll-derivedtargetProgress, not synthetic chapter autoplay. Reverse immediately, settle exactly, bypass on reduced-motion/direct editing/touch where appropriate. Read momentum. - Adapt mobile composition separately; do not simply shrink a desktop window. Protect accepted desktop imagery and local chapter mapping.
Budgets, loading and quality gates
- Obtain Owner-approved total frame/byte/time limits before new media production; apply aggregate size across responsive tiers. Never invent the platform's remaining Work quota. See media budget, work budget, and loading/frame gaps.
- For substantial work, checkpoint after meaningful phases, SHA-256 each snapshot, and ensure the owner can really retrieve it; local files alone may vanish. See checkpoint protocol.
- Preserve image bytes; media readiness,
img.decode, poster/fallback, reverse scroll, frame cadence, no horizontal overflow, caption collisions, alt/semantics andprefers-reduced-motionneed browser checks. See browser evidence, semantic hero and embed/staging. - Distinguish source test, numeric test, local browser, hosted browser, and real device. Never convert a lower evidence tier into a live/physical-device PASS.
Use packaged scripts when relevant
Execute scripts explicitly from this skill's scripts/ folder; optional --report/--out outputs require authorized workspaces:
validate_visual_timeline.py,validate_text_overlays.py,audit_caption_coverage.py,audit_hero_html.py,audit_scroll_momentum.py— structural/log preflights, not visual acceptance.media_budget_contract.py,configure_owner_budget.py— planned/actual media caps;work_budget_guard.py,checkpoint_manager.py— local budget/recovery ledgers.validate_six_frame_proof.py— only for explicitly approved 3D-source profile, not image-first.- Examples in
assets/are SYNTHETIC or NOT OWNER APPROVED, never deployment defaults.
Delivery
Provide source identity/hash, selected profile, scene/element timeline, before/after images when authorized, mobile and desktop evidence, measured versus inferred findings, budget/checkpoint receipt, accessible content and a table of PASS / FAIL / NOT TESTED / BLOCKED. Stop at the authorized gate; deployment requires separate explicit approval.