← PicsartCONTENT HISTORY
WHAT CHANGED
changed
Update to Picsart
Snapshot Nov 7, 2026 · 00:09 UTC · version 2.4.0
Collection source: downloaded plugin package.
Instructions updated for motion-intent
Skill instructions — Now says: ported-from: motion-intent 1.2 at 55.75.0
Observed in instructions or declared skills. Runtime behavior has not been tested.
Save Picsart and check future changes in Following.
Up to 3 plugins free · no card required.
All fields and original values.
Skill instructions
Before
Nov 3, 2026 · 18:14 UTC
No removed lines (ordering may have changed)
After
Nov 7, 2026 · 00:09 UTC
ported-from: motion-intent 1.2 at 55.75.0
Compare saved observations
Download comparison JSONSKILL.md text changes
− removed · + added
--- before +++ after @@ -4,6 +4,7 @@ metadata: category: authoring version: '1.2' + ported-from: motion-intent 1.2 at 55.75.0 --- # Motion by intent
Full technical diff · 1 changed fields
changed /skill_md_contents
BEFORE
"---\nname: motion-intent\ndescription: Choose and diagnose scene motion from communication intent before selecting presets or keyframes. Use when a user says “make this motion feel…”, asks to choreograph a scene, choose timing/easing/stagger, establish motion hierarchy, or explain why an animation feels flat, noisy, robotic, sluggish, or unfinished.\nmetadata:\n category: authoring\n version: '1.2'\n---\n\n# Motion by intent\n\nChoose motion from communication intent before choosing presets or keyframes. Motion spends\nattention; an element should move because the beat needs it, not merely because it can.\n\nThis recipe uses three evidence labels:\n\n- **scene fact** — behavior defined by the format, validator, resolver, or authoring tools.\n- **Motion heuristic** — a useful starting rule that must be judged in the actual scene.\n- **Project choice** — a local style decision that should stay consistent inside one piece.\n\nDo not present a heuristic or project choice as something scene enforces. Exact starting values\nbelow are trials, not universal constants.\n\n## Decide in this order\n\n1. **Communication goal** — what must the viewer understand or notice?\n2. **Visual leader** — which one element owns the beat, and what stays subordinate?\n3. **Spatial relationship** — reveal, replace, connect, emphasize, orient, or resolve.\n4. **Timing and sequence** — what precedes, overlaps, follows, and holds?\n5. **Motion character** — direct, soft, restrained, playful, urgent, or continuous.\n6. **Settle and readable hold** — what must become still, and for how long?\n7. **Evaluation** — did the result communicate the original goal without added competition?\n\nOnly after these decisions should an author choose scene properties, presets, or keyframes.\nConcrete syntax is `picsart_scene_recipe scene-animation-reference`; reusable realizations are\n`picsart_scene_recipe motion-pattern-library`. When the question spans consecutive beats or a scene boundary,\nroute the relationship and seam decision to `picsart_scene_recipe motion-scene-continuity`.\n\n## 1. Communication goal\n\n| Goal | Communication job | Possible scene realization |\n|---|---|---|\n| **Reveal** | introduce content | an entrance preset, text animation, or explicit opacity/transform track |\n| **Conceal** | remove content | an exit preset or explicit opacity/transform track |\n| **Replace** | make one state become another | keep shared layers continuous; remove and introduce only what changes |\n| **Connect** | show a relationship or journey | position path, parenting, anchor and rotation, or a drawn path |\n| **Emphasize** | redirect attention mid-beat | a restrained scale, position, rotation, focus, or colour change |\n| **Orient** | communicate direction, order, or progress | coordinated position tracks or a catalog transition such as `slide`, `diagonal_slide` or `page_curl` |\n| **Resolve** | show that an action reached its intended result | return to rest, reduce amplitude, then hold if comprehension needs it |\n| **Atmosphere** | keep an otherwise static interval alive | low-amplitude continuous motion that never becomes the leader |\n\n**Motion heuristic:** one verb is usually enough for one beat. If the goal needs “reveal, emphasize,\nspin, and celebrate,” split it into beats or choose the one communication job that matters.\n\n## 2. Visual leader\n\nName the leader before assigning motion. Supporting elements may prepare for it, follow it, or remain\nstill. A useful hierarchy trial is:\n\n| Role | Starting treatment |\n|---|---|\n| leader | largest or clearest change in the beat |\n| support | smaller change, later start, or fewer animated properties |\n| ambient | slow, low-amplitude continuity, or no motion |\n\nThese are **motion heuristics**, not amplitude requirements. A leader can also win through isolation,\ncontrast, or timing while moving less than its support.\n\nFor repeated items, decide whether the group is one idea or a reading order. Move the group together\nfor one idea; stagger its children when order itself communicates something.\n\n## 3. Spatial relationship and property\n\n**scene fact:** layer Placement can animate `position`, `scale`, `rotation`, `opacity`, and\n`anchor`; audio can animate `volume`. Transform components add independently editable motion outside\nPlacement. Exact catalogs and content-kind restrictions live in `scene-animation-reference`.\n\nChoose the minimum property set that carries the relationship:\n\n| Property | Often communicates | Context warning |\n|---|---|---|\n| `position` | arrival, departure, direction, relationship | direction has cultural and layout context; right is not universally “progress” |\n| `scale` | importance, proximity, emergence | large deltas can feel like a camera move or a toy-like bounce |\n| `rotation` | redirection, process, instability | continuous rotation implies an ongoing state rather than a resolved beat |\n| `opacity` | presence or continuity | opacity alone may be exactly right for a quiet reveal; pair it only when another property adds meaning |\n| `anchor` | the point a scale or rotation acts around | animate it only when changing the pivot is itself useful |\n\n**Motion heuristic:** one or two coordinated properties are easier to read than several unrelated\nones. When opacity and a transform describe one arrival, try ending them together first; offset them\nonly for a deliberate two-stage read.\n\n### Shared-element continuity\n\nFor an element that exists in consecutive arrangements, first try keeping one layer on screen and\nanimating its delta. This often reads as continuity better than cross-fading two copies. A cross-fade\nremains valid when identity is intentionally ambiguous, continuity is unimportant, or the states\ncannot share one authored layer.\n\n## 4. Timing and sequence\n\nscene defines time values; it does not define an ideal pace. These are for ambient or emotional\npacing; a UI element move (a button, a card, a list row) stays at motion-design-craft's 200–400 ms. Build a small **project-choice**\npalette from the brief, then test it in context. These are example starting palettes:\n\n| Character | quick | standard | slow | Overshoot trial |\n|---|---:|---:|---:|---:|\n| calm / editorial | 0.5s | 0.8s | 1.6s | none |\n| restrained / professional | 0.3s | 0.5s | 0.9s | none to about 3% |\n| playful | 0.25s | 0.4s | 0.7s | about 10–20% on the leader only |\n| energetic | 0.15s | 0.3s | 0.5s | about 15–30% when readability survives |\n\nThe names and values above are not format registers. Replace them when the project's pacing,\nframe rate, travel, density, or audience calls for something else.\n\nUseful **heuristic trials**:\n\n- Longer travel often needs more time; preview rather than applying a fixed distance multiplier.\n- Try exits shorter than entrances when removal should spend less attention, but keep them long\n enough to remain legible.\n- For siblings, start with roughly `0.05–0.12s` between items and keep the whole cascade compact;\n compress or expand it to match density and reading order.\n- Very short spatial moves can read as pops at common video frame rates. Around `0.15s` is a useful\n warning threshold, not a minimum enforced by scene.\n- A hold should cover comprehension or reading, not satisfy a fixed ratio to motion duration.\n\nWrite the chosen palette beside the scene as a project decision. Consistent values can create a\nregister, but variation is justified when distance, hierarchy, or meaning changes.\n\n## 5. Motion character and easing\n\n**scene fact:** if an animatable keyframe segment omits easing, interpolation is linear. State\neasing explicitly whenever constant-rate motion is not intended.\n\n**Motion heuristic:** purpose and continuity choose easing; direction is only a starting clue.\n\n| Desired read | First easing to try | Legitimate exception |\n|---|---|---|\n| arrive and settle | `ease_out` | `linear` for mechanical or intentionally constant motion |\n| leave and accelerate | `ease_in` | `ease_out` for a release that starts abruptly and softens |\n| reposition while remaining visible | `ease_in_out` | `linear` for a camera glide, progress, or conveyor-like motion |\n| switch instantly | `step` | use a short eased change when continuity matters |\n\nOvershoot is a design choice, not a separate scene semantic. Start with none, add it only to the\nleader when the requested character needs it, and judge the rendered peak rather than equating a\nBézier control ordinate with an overshoot percentage. A single overshoot and a damped multi-cycle\nspring communicate different things; patterns for both are in `motion-pattern-library`.\n\n## 6. Settle and readable hold\n\nAnticipation, action, reaction, and resolution are optional choreography tools—not mandatory phases.\nA practical sequence is:\n\n1. anticipate only when it clarifies a strong directional action;\n2. perform the action;\n3. let a subordinate react only if the reaction adds meaning;\n4. resolve or intentionally cut away;\n5. hold only as long as recognition or reading needs.\n\nStarting with `0.1–0.2s` of stability can reveal whether a beat felt unfinished, but the correct hold\ndepends on frame rate, edit rhythm, and content. A hard cut at the action endpoint can also be the\nintended punctuation.\n\n## 7. Evaluate against the goal\n\nTreat perceptual diagnoses as hypotheses and change one variable at a time.\n\n| Question | Evidence available here |\n|---|---|\n| Did geometry and opacity follow the intended beat? | `picsart_scene_query_layout` at events, extrema and through the hold, compared with independent expected values |\n| Did supported relationships hold at the samples? | `picsart_scene_lint_layout` with scoped targets, explicit windows and supported relationships; inspect every verdict |\n| Did painted objects move as intended? | Inspect continuous exported video; painted motion is not tracked automatically (`picsart_scene_check_motion` covers evaluated geometry) |\n| Is the document structurally valid? | `picsart_scene_validate` after edits |\n| Does the beat read at viewing speed? | Watch it in sequence; unexecuted numerical checks remain unchecked |\n\nLayout and validation can prove authored structure; they cannot prove that motion feels confident,\nplayful, elegant, or restrained. Those remain contextual judgments.\n\nBefore authoring, write a short acceptance card: **goal, target IDs/scopes, root-clock events,\nexpected outcome, tolerance, sample schedule, numerical check, visual check**. Pick the matching\npattern from `picsart_scene_recipe motion-pattern-library` and fetch its **Validation by pattern** reference;\nit maps each construction to existing tools\nand states what they cannot establish. Set targets from the design, not from the output being tested.\nAfter execution, record measured / visually inspected / unchecked evidence separately. Fix a failed\nbeat and rerun its window and adjacent seams; do not widen tolerances merely to obtain a pass.\n\n## Before calling a beat done\n\n- [ ] The communication goal is one sentence, not an effect name.\n- [ ] One leader is named; support and ambient motion are intentional, including intentional stillness.\n- [ ] The property choice communicates the relationship.\n- [ ] Numerical values are identified as project choices or heuristic trials unless the format defines them.\n- [ ] Easing is explicit unless constant-rate motion is intended.\n- [ ] The beat resolves or cuts intentionally, with enough reading time for its content.\n- [ ] Validation/layout evidence does not prove a subjective feeling; describe the actual visual observation separately.\n- [ ] The selected pattern's numerical and visual checks ran, or the missing evidence is explicitly listed.\n\nEmotion-to-motion starting points and symptom diagnosis are in\n[references/emotion-and-diagnosis.md](references/emotion-and-diagnosis.md).\n\n---\n\nThe existing goal/register/emotion framing was adapted for video from LottieFiles' `motion-design`\nskill (MIT © 2025 LottieFiles) and re-derived in scene vocabulary; that attribution remains for\nthe material already derived from it.\n\n**Research note:** the evidence-boundary and choreography guidance added here is independently\nauthored for scene from general motion-design principles and local evaluation. Hyperframes was\nreviewed as research inspiration; its wording, code, examples, and organization were not adopted.\n"
AFTER
"---\nname: motion-intent\ndescription: Choose and diagnose scene motion from communication intent before selecting presets or keyframes. Use when a user says “make this motion feel…”, asks to choreograph a scene, choose timing/easing/stagger, establish motion hierarchy, or explain why an animation feels flat, noisy, robotic, sluggish, or unfinished.\nmetadata:\n category: authoring\n version: '1.2'\n ported-from: motion-intent 1.2 at 55.75.0\n---\n\n# Motion by intent\n\nChoose motion from communication intent before choosing presets or keyframes. Motion spends\nattention; an element should move because the beat needs it, not merely because it can.\n\nThis recipe uses three evidence labels:\n\n- **scene fact** — behavior defined by the format, validator, resolver, or authoring tools.\n- **Motion heuristic** — a useful starting rule that must be judged in the actual scene.\n- **Project choice** — a local style decision that should stay consistent inside one piece.\n\nDo not present a heuristic or project choice as something scene enforces. Exact starting values\nbelow are trials, not universal constants.\n\n## Decide in this order\n\n1. **Communication goal** — what must the viewer understand or notice?\n2. **Visual leader** — which one element owns the beat, and what stays subordinate?\n3. **Spatial relationship** — reveal, replace, connect, emphasize, orient, or resolve.\n4. **Timing and sequence** — what precedes, overlaps, follows, and holds?\n5. **Motion character** — direct, soft, restrained, playful, urgent, or continuous.\n6. **Settle and readable hold** — what must become still, and for how long?\n7. **Evaluation** — did the result communicate the original goal without added competition?\n\nOnly after these decisions should an author choose scene properties, presets, or keyframes.\nConcrete syntax is `picsart_scene_recipe scene-animation-reference`; reusable realizations are\n`picsart_scene_recipe motion-pattern-library`. When the question spans consecutive beats or a scene boundary,\nroute the relationship and seam decision to `picsart_scene_recipe motion-scene-continuity`.\n\n## 1. Communication goal\n\n| Goal | Communication job | Possible scene realization |\n|---|---|---|\n| **Reveal** | introduce content | an entrance preset, text animation, or explicit opacity/transform track |\n| **Conceal** | remove content | an exit preset or explicit opacity/transform track |\n| **Replace** | make one state become another | keep shared layers continuous; remove and introduce only what changes |\n| **Connect** | show a relationship or journey | position path, parenting, anchor and rotation, or a drawn path |\n| **Emphasize** | redirect attention mid-beat | a restrained scale, position, rotation, focus, or colour change |\n| **Orient** | communicate direction, order, or progress | coordinated position tracks or a catalog transition such as `slide`, `diagonal_slide` or `page_curl` |\n| **Resolve** | show that an action reached its intended result | return to rest, reduce amplitude, then hold if comprehension needs it |\n| **Atmosphere** | keep an otherwise static interval alive | low-amplitude continuous motion that never becomes the leader |\n\n**Motion heuristic:** one verb is usually enough for one beat. If the goal needs “reveal, emphasize,\nspin, and celebrate,” split it into beats or choose the one communication job that matters.\n\n## 2. Visual leader\n\nName the leader before assigning motion. Supporting elements may prepare for it, follow it, or remain\nstill. A useful hierarchy trial is:\n\n| Role | Starting treatment |\n|---|---|\n| leader | largest or clearest change in the beat |\n| support | smaller change, later start, or fewer animated properties |\n| ambient | slow, low-amplitude continuity, or no motion |\n\nThese are **motion heuristics**, not amplitude requirements. A leader can also win through isolation,\ncontrast, or timing while moving less than its support.\n\nFor repeated items, decide whether the group is one idea or a reading order. Move the group together\nfor one idea; stagger its children when order itself communicates something.\n\n## 3. Spatial relationship and property\n\n**scene fact:** layer Placement can animate `position`, `scale`, `rotation`, `opacity`, and\n`anchor`; audio can animate `volume`. Transform components add independently editable motion outside\nPlacement. Exact catalogs and content-kind restrictions live in `scene-animation-reference`.\n\nChoose the minimum property set that carries the relationship:\n\n| Property | Often communicates | Context warning |\n|---|---|---|\n| `position` | arrival, departure, direction, relationship | direction has cultural and layout context; right is not universally “progress” |\n| `scale` | importance, proximity, emergence | large deltas can feel like a camera move or a toy-like bounce |\n| `rotation` | redirection, process, instability | continuous rotation implies an ongoing state rather than a resolved beat |\n| `opacity` | presence or continuity | opacity alone may be exactly right for a quiet reveal; pair it only when another property adds meaning |\n| `anchor` | the point a scale or rotation acts around | animate it only when changing the pivot is itself useful |\n\n**Motion heuristic:** one or two coordinated properties are easier to read than several unrelated\nones. When opacity and a transform describe one arrival, try ending them together first; offset them\nonly for a deliberate two-stage read.\n\n### Shared-element continuity\n\nFor an element that exists in consecutive arrangements, first try keeping one layer on screen and\nanimating its delta. This often reads as continuity better than cross-fading two copies. A cross-fade\nremains valid when identity is intentionally ambiguous, continuity is unimportant, or the states\ncannot share one authored layer.\n\n## 4. Timing and sequence\n\nscene defines time values; it does not define an ideal pace. These are for ambient or emotional\npacing; a UI element move (a button, a card, a list row) stays at motion-design-craft's 200–400 ms. Build a small **project-choice**\npalette from the brief, then test it in context. These are example starting palettes:\n\n| Character | quick | standard | slow | Overshoot trial |\n|---|---:|---:|---:|---:|\n| calm / editorial | 0.5s | 0.8s | 1.6s | none |\n| restrained / professional | 0.3s | 0.5s | 0.9s | none to about 3% |\n| playful | 0.25s | 0.4s | 0.7s | about 10–20% on the leader only |\n| energetic | 0.15s | 0.3s | 0.5s | about 15–30% when readability survives |\n\nThe names and values above are not format registers. Replace them when the project's pacing,\nframe rate, travel, density, or audience calls for something else.\n\nUseful **heuristic trials**:\n\n- Longer travel often needs more time; preview rather than applying a fixed distance multiplier.\n- Try exits shorter than entrances when removal should spend less attention, but keep them long\n enough to remain legible.\n- For siblings, start with roughly `0.05–0.12s` between items and keep the whole cascade compact;\n compress or expand it to match density and reading order.\n- Very short spatial moves can read as pops at common video frame rates. Around `0.15s` is a useful\n warning threshold, not a minimum enforced by scene.\n- A hold should cover comprehension or reading, not satisfy a fixed ratio to motion duration.\n\nWrite the chosen palette beside the scene as a project decision. Consistent values can create a\nregister, but variation is justified when distance, hierarchy, or meaning changes.\n\n## 5. Motion character and easing\n\n**scene fact:** if an animatable keyframe segment omits easing, interpolation is linear. State\neasing explicitly whenever constant-rate motion is not intended.\n\n**Motion heuristic:** purpose and continuity choose easing; direction is only a starting clue.\n\n| Desired read | First easing to try | Legitimate exception |\n|---|---|---|\n| arrive and settle | `ease_out` | `linear` for mechanical or intentionally constant motion |\n| leave and accelerate | `ease_in` | `ease_out` for a release that starts abruptly and softens |\n| reposition while remaining visible | `ease_in_out` | `linear` for a camera glide, progress, or conveyor-like motion |\n| switch instantly | `step` | use a short eased change when continuity matters |\n\nOvershoot is a design choice, not a separate scene semantic. Start with none, add it only to the\nleader when the requested character needs it, and judge the rendered peak rather than equating a\nBézier control ordinate with an overshoot percentage. A single overshoot and a damped multi-cycle\nspring communicate different things; patterns for both are in `motion-pattern-library`.\n\n## 6. Settle and readable hold\n\nAnticipation, action, reaction, and resolution are optional choreography tools—not mandatory phases.\nA practical sequence is:\n\n1. anticipate only when it clarifies a strong directional action;\n2. perform the action;\n3. let a subordinate react only if the reaction adds meaning;\n4. resolve or intentionally cut away;\n5. hold only as long as recognition or reading needs.\n\nStarting with `0.1–0.2s` of stability can reveal whether a beat felt unfinished, but the correct hold\ndepends on frame rate, edit rhythm, and content. A hard cut at the action endpoint can also be the\nintended punctuation.\n\n## 7. Evaluate against the goal\n\nTreat perceptual diagnoses as hypotheses and change one variable at a time.\n\n| Question | Evidence available here |\n|---|---|\n| Did geometry and opacity follow the intended beat? | `picsart_scene_query_layout` at events, extrema and through the hold, compared with independent expected values |\n| Did supported relationships hold at the samples? | `picsart_scene_lint_layout` with scoped targets, explicit windows and supported relationships; inspect every verdict |\n| Did painted objects move as intended? | Inspect continuous exported video; painted motion is not tracked automatically (`picsart_scene_check_motion` covers evaluated geometry) |\n| Is the document structurally valid? | `picsart_scene_validate` after edits |\n| Does the beat read at viewing speed? | Watch it in sequence; unexecuted numerical checks remain unchecked |\n\nLayout and validation can prove authored structure; they cannot prove that motion feels confident,\nplayful, elegant, or restrained. Those remain contextual judgments.\n\nBefore authoring, write a short acceptance card: **goal, target IDs/scopes, root-clock events,\nexpected outcome, tolerance, sample schedule, numerical check, visual check**. Pick the matching\npattern from `picsart_scene_recipe motion-pattern-library` and fetch its **Validation by pattern** reference;\nit maps each construction to existing tools\nand states what they cannot establish. Set targets from the design, not from the output being tested.\nAfter execution, record measured / visually inspected / unchecked evidence separately. Fix a failed\nbeat and rerun its window and adjacent seams; do not widen tolerances merely to obtain a pass.\n\n## Before calling a beat done\n\n- [ ] The communication goal is one sentence, not an effect name.\n- [ ] One leader is named; support and ambient motion are intentional, including intentional stillness.\n- [ ] The property choice communicates the relationship.\n- [ ] Numerical values are identified as project choices or heuristic trials unless the format defines them.\n- [ ] Easing is explicit unless constant-rate motion is intended.\n- [ ] The beat resolves or cuts intentionally, with enough reading time for its content.\n- [ ] Validation/layout evidence does not prove a subjective feeling; describe the actual visual observation separately.\n- [ ] The selected pattern's numerical and visual checks ran, or the missing evidence is explicitly listed.\n\nEmotion-to-motion starting points and symptom diagnosis are in\n[references/emotion-and-diagnosis.md](references/emotion-and-diagnosis.md).\n\n---\n\nThe existing goal/register/emotion framing was adapted for video from LottieFiles' `motion-design`\nskill (MIT © 2025 LottieFiles) and re-derived in scene vocabulary; that attribution remains for\nthe material already derived from it.\n\n**Research note:** the evidence-boundary and choreography guidance added here is independently\nauthored for scene from general motion-design principles and local evaluation. Hyperframes was\nreviewed as research inspiration; its wording, code, examples, and organization were not adopted.\n"
Full snapshot data
{
"description": "Choose and diagnose scene motion from communication intent before selecting presets or keyframes. Use when a user says “make this motion feel…”, asks to choreograph a scene, choose timing/easing/stagger, establish motion hierarchy, or explain why an animation feels flat, noisy, robotic, sluggish, or unfinished.",
"included_files": [
{
"relative_path": "references/emotion-and-diagnosis.md",
"size_in_bytes": 5612
}
],
"name": "motion-intent",
"skill_md_contents": "---\nname: motion-intent\ndescription: Choose and diagnose scene motion from communication intent before selecting presets or keyframes. Use when a user says “make this motion feel…”, asks to choreograph a scene, choose timing/easing/stagger, establish motion hierarchy, or explain why an animation feels flat, noisy, robotic, sluggish, or unfinished.\nmetadata:\n category: authoring\n version: '1.2'\n ported-from: motion-intent 1.2 at 55.75.0\n---\n\n# Motion by intent\n\nChoose motion from communication intent before choosing presets or keyframes. Motion spends\nattention; an element should move because the beat needs it, not merely because it can.\n\nThis recipe uses three evidence labels:\n\n- **scene fact** — behavior defined by the format, validator, resolver, or authoring tools.\n- **Motion heuristic** — a useful starting rule that must be judged in the actual scene.\n- **Project choice** — a local style decision that should stay consistent inside one piece.\n\nDo not present a heuristic or project choice as something scene enforces. Exact starting values\nbelow are trials, not universal constants.\n\n## Decide in this order\n\n1. **Communication goal** — what must the viewer understand or notice?\n2. **Visual leader** — which one element owns the beat, and what stays subordinate?\n3. **Spatial relationship** — reveal, replace, connect, emphasize, orient, or resolve.\n4. **Timing and sequence** — what precedes, overlaps, follows, and holds?\n5. **Motion character** — direct, soft, restrained, playful, urgent, or continuous.\n6. **Settle and readable hold** — what must become still, and for how long?\n7. **Evaluation** — did the result communicate the original goal without added competition?\n\nOnly after these decisions should an author choose scene properties, presets, or keyframes.\nConcrete syntax is `picsart_scene_recipe scene-animation-reference`; reusable realizations are\n`picsart_scene_recipe motion-pattern-library`. When the question spans consecutive beats or a scene boundary,\nroute the relationship and seam decision to `picsart_scene_recipe motion-scene-continuity`.\n\n## 1. Communication goal\n\n| Goal | Communication job | Possible scene realization |\n|---|---|---|\n| **Reveal** | introduce content | an entrance preset, text animation, or explicit opacity/transform track |\n| **Conceal** | remove content | an exit preset or explicit opacity/transform track |\n| **Replace** | make one state become another | keep shared layers continuous; remove and introduce only what changes |\n| **Connect** | show a relationship or journey | position path, parenting, anchor and rotation, or a drawn path |\n| **Emphasize** | redirect attention mid-beat | a restrained scale, position, rotation, focus, or colour change |\n| **Orient** | communicate direction, order, or progress | coordinated position tracks or a catalog transition such as `slide`, `diagonal_slide` or `page_curl` |\n| **Resolve** | show that an action reached its intended result | return to rest, reduce amplitude, then hold if comprehension needs it |\n| **Atmosphere** | keep an otherwise static interval alive | low-amplitude continuous motion that never becomes the leader |\n\n**Motion heuristic:** one verb is usually enough for one beat. If the goal needs “reveal, emphasize,\nspin, and celebrate,” split it into beats or choose the one communication job that matters.\n\n## 2. Visual leader\n\nName the leader before assigning motion. Supporting elements may prepare for it, follow it, or remain\nstill. A useful hierarchy trial is:\n\n| Role | Starting treatment |\n|---|---|\n| leader | largest or clearest change in the beat |\n| support | smaller change, later start, or fewer animated properties |\n| ambient | slow, low-amplitude continuity, or no motion |\n\nThese are **motion heuristics**, not amplitude requirements. A leader can also win through isolation,\ncontrast, or timing while moving less than its support.\n\nFor repeated items, decide whether the group is one idea or a reading order. Move the group together\nfor one idea; stagger its children when order itself communicates something.\n\n## 3. Spatial relationship and property\n\n**scene fact:** layer Placement can animate `position`, `scale`, `rotation`, `opacity`, and\n`anchor`; audio can animate `volume`. Transform components add independently editable motion outside\nPlacement. Exact catalogs and content-kind restrictions live in `scene-animation-reference`.\n\nChoose the minimum property set that carries the relationship:\n\n| Property | Often communicates | Context warning |\n|---|---|---|\n| `position` | arrival, departure, direction, relationship | direction has cultural and layout context; right is not universally “progress” |\n| `scale` | importance, proximity, emergence | large deltas can feel like a camera move or a toy-like bounce |\n| `rotation` | redirection, process, instability | continuous rotation implies an ongoing state rather than a resolved beat |\n| `opacity` | presence or continuity | opacity alone may be exactly right for a quiet reveal; pair it only when another property adds meaning |\n| `anchor` | the point a scale or rotation acts around | animate it only when changing the pivot is itself useful |\n\n**Motion heuristic:** one or two coordinated properties are easier to read than several unrelated\nones. When opacity and a transform describe one arrival, try ending them together first; offset them\nonly for a deliberate two-stage read.\n\n### Shared-element continuity\n\nFor an element that exists in consecutive arrangements, first try keeping one layer on screen and\nanimating its delta. This often reads as continuity better than cross-fading two copies. A cross-fade\nremains valid when identity is intentionally ambiguous, continuity is unimportant, or the states\ncannot share one authored layer.\n\n## 4. Timing and sequence\n\nscene defines time values; it does not define an ideal pace. These are for ambient or emotional\npacing; a UI element move (a button, a card, a list row) stays at motion-design-craft's 200–400 ms. Build a small **project-choice**\npalette from the brief, then test it in context. These are example starting palettes:\n\n| Character | quick | standard | slow | Overshoot trial |\n|---|---:|---:|---:|---:|\n| calm / editorial | 0.5s | 0.8s | 1.6s | none |\n| restrained / professional | 0.3s | 0.5s | 0.9s | none to about 3% |\n| playful | 0.25s | 0.4s | 0.7s | about 10–20% on the leader only |\n| energetic | 0.15s | 0.3s | 0.5s | about 15–30% when readability survives |\n\nThe names and values above are not format registers. Replace them when the project's pacing,\nframe rate, travel, density, or audience calls for something else.\n\nUseful **heuristic trials**:\n\n- Longer travel often needs more time; preview rather than applying a fixed distance multiplier.\n- Try exits shorter than entrances when removal should spend less attention, but keep them long\n enough to remain legible.\n- For siblings, start with roughly `0.05–0.12s` between items and keep the whole cascade compact;\n compress or expand it to match density and reading order.\n- Very short spatial moves can read as pops at common video frame rates. Around `0.15s` is a useful\n warning threshold, not a minimum enforced by scene.\n- A hold should cover comprehension or reading, not satisfy a fixed ratio to motion duration.\n\nWrite the chosen palette beside the scene as a project decision. Consistent values can create a\nregister, but variation is justified when distance, hierarchy, or meaning changes.\n\n## 5. Motion character and easing\n\n**scene fact:** if an animatable keyframe segment omits easing, interpolation is linear. State\neasing explicitly whenever constant-rate motion is not intended.\n\n**Motion heuristic:** purpose and continuity choose easing; direction is only a starting clue.\n\n| Desired read | First easing to try | Legitimate exception |\n|---|---|---|\n| arrive and settle | `ease_out` | `linear` for mechanical or intentionally constant motion |\n| leave and accelerate | `ease_in` | `ease_out` for a release that starts abruptly and softens |\n| reposition while remaining visible | `ease_in_out` | `linear` for a camera glide, progress, or conveyor-like motion |\n| switch instantly | `step` | use a short eased change when continuity matters |\n\nOvershoot is a design choice, not a separate scene semantic. Start with none, add it only to the\nleader when the requested character needs it, and judge the rendered peak rather than equating a\nBézier control ordinate with an overshoot percentage. A single overshoot and a damped multi-cycle\nspring communicate different things; patterns for both are in `motion-pattern-library`.\n\n## 6. Settle and readable hold\n\nAnticipation, action, reaction, and resolution are optional choreography tools—not mandatory phases.\nA practical sequence is:\n\n1. anticipate only when it clarifies a strong directional action;\n2. perform the action;\n3. let a subordinate react only if the reaction adds meaning;\n4. resolve or intentionally cut away;\n5. hold only as long as recognition or reading needs.\n\nStarting with `0.1–0.2s` of stability can reveal whether a beat felt unfinished, but the correct hold\ndepends on frame rate, edit rhythm, and content. A hard cut at the action endpoint can also be the\nintended punctuation.\n\n## 7. Evaluate against the goal\n\nTreat perceptual diagnoses as hypotheses and change one variable at a time.\n\n| Question | Evidence available here |\n|---|---|\n| Did geometry and opacity follow the intended beat? | `picsart_scene_query_layout` at events, extrema and through the hold, compared with independent expected values |\n| Did supported relationships hold at the samples? | `picsart_scene_lint_layout` with scoped targets, explicit windows and supported relationships; inspect every verdict |\n| Did painted objects move as intended? | Inspect continuous exported video; painted motion is not tracked automatically (`picsart_scene_check_motion` covers evaluated geometry) |\n| Is the document structurally valid? | `picsart_scene_validate` after edits |\n| Does the beat read at viewing speed? | Watch it in sequence; unexecuted numerical checks remain unchecked |\n\nLayout and validation can prove authored structure; they cannot prove that motion feels confident,\nplayful, elegant, or restrained. Those remain contextual judgments.\n\nBefore authoring, write a short acceptance card: **goal, target IDs/scopes, root-clock events,\nexpected outcome, tolerance, sample schedule, numerical check, visual check**. Pick the matching\npattern from `picsart_scene_recipe motion-pattern-library` and fetch its **Validation by pattern** reference;\nit maps each construction to existing tools\nand states what they cannot establish. Set targets from the design, not from the output being tested.\nAfter execution, record measured / visually inspected / unchecked evidence separately. Fix a failed\nbeat and rerun its window and adjacent seams; do not widen tolerances merely to obtain a pass.\n\n## Before calling a beat done\n\n- [ ] The communication goal is one sentence, not an effect name.\n- [ ] One leader is named; support and ambient motion are intentional, including intentional stillness.\n- [ ] The property choice communicates the relationship.\n- [ ] Numerical values are identified as project choices or heuristic trials unless the format defines them.\n- [ ] Easing is explicit unless constant-rate motion is intended.\n- [ ] The beat resolves or cuts intentionally, with enough reading time for its content.\n- [ ] Validation/layout evidence does not prove a subjective feeling; describe the actual visual observation separately.\n- [ ] The selected pattern's numerical and visual checks ran, or the missing evidence is explicitly listed.\n\nEmotion-to-motion starting points and symptom diagnosis are in\n[references/emotion-and-diagnosis.md](references/emotion-and-diagnosis.md).\n\n---\n\nThe existing goal/register/emotion framing was adapted for video from LottieFiles' `motion-design`\nskill (MIT © 2025 LottieFiles) and re-derived in scene vocabulary; that attribution remains for\nthe material already derived from it.\n\n**Research note:** the evidence-boundary and choreography guidance added here is independently\nauthored for scene from general motion-design principles and local evaluation. Hyperframes was\nreviewed as research inspiration; its wording, code, examples, and organization were not adopted.\n"
}SHA-256 of public snapshot: 6dbe1d4f2173c2f715a3d2b8b0bee2bd304e6d8a6cb917619d7be1ab2901ca1c