@docyrus/animaguru 0.16.1 → 0.17.0

package/skills/animaguru-art-motion/references/art/greek-vase-painting.md

@@ -0,0 +1,27 @@
+# Greek black-figure pottery
+
+Terracotta, black silhouettes and incised internal detail, suited to friezes and mythic sequences.
+
+## Material and construction
+
+Lay a terracotta field over subtle fired-clay variation. Use black-figure silhouettes, pale incised lines and restrained maroon/white accents. Assemble meanders as repeatable border tiles. Internal incisions must survive the final viewing size, especially where dark forms overlap.
+
+## Palette
+
+Terracotta orange, near-black, cream and muted wine red.
+
+## Motion
+
+Counter-scroll the upper and lower meanders, let a ship travel across the frieze, and use a dolphin leap or lyre pluck as the main motif. Keep figures on a shared register rather than adding modern perspective.
+
+## Transition
+
+Roll the pottery surface or carry the next scene in a moving frieze.
+
+## Failure cases and adaptation
+
+Overlapping black bodies merge unless incised separation lines remain visible. Match the repeat period to the actual meander tile. A smooth shaded body looks like a contemporary illustration rather than ceramic decoration.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/03_greek.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/hopper-hard-light.md

@@ -0,0 +1,27 @@
+# Quiet realism and hard sunlight
+
+Architectural stillness, warm sunlight and sharply bounded cool shadow.
+
+## Material and construction
+
+Cache complete lit and shaded versions, then reveal the lit version with a moving hard-edged light polygon. Include the subject in both treatments so light really crosses the face and clothes. Shadows follow the current pose. Add low-contrast vertical oil texture rather than obvious marbling.
+
+## Palette
+
+Cool shade around #5D8679 and floor #5B3D2B; sunlight #EFCF8F and floor #C98B52.
+
+## Motion
+
+A sunlight patch crosses the room, a cord sways and a few dust particles appear only within the beam. Let an understated human action carry the scene. The source used about 95 px of light travel in 1.2 s as a starting point.
+
+## Transition
+
+A hard sunshine edge reveals the next space.
+
+## Failure cases and adaptation
+
+A fixed overlay on an unlit character looks pasted on. Strong horizontal texture reads as marbling; a quarter-resolution soft-light texture near .2 alpha was subtler than .35. Keep the subject readable in shadow; stillness is a valid part of this treatment.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/24_hopper.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/impressionist-dabs.md

@@ -0,0 +1,27 @@
+# Impressionist dabs
+
+Broken light and short varied marks, distinct from uniform pointillist dots.
+
+## Material and construction
+
+Start from a readable color/value composition. Repaint regions with short ellipses and dabs sampled from local warm/cool palettes. Keep stroke positions seeded; vary their light or orientation smoothly. Use a finer separate pass for window frames, hands and facial landmarks. Avoid thick outlined strokes.
+
+## Palette
+
+Pastel regional palettes with warm sunlight and cool shadows; sample several colors per region rather than adding RGB noise to one base color.
+
+## Motion
+
+Dappled light crosses a surface, poppies or grass sway, clouds drift and a ribbon catches the breeze. Select a main action and supporting loops that fit the scene.
+
+## Transition
+
+Build the next image from dabs or let a field of marks carry the cut.
+
+## Failure cases and adaptation
+
+Uniform pure-color equal-size dots belong to pointillism. Whole-frame reseeding reads as flashing. Faces and hands were weak in the source demo: protect their contours and inspect a full-size still before adding more texture.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/08_impressionism.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/ink-wash.md

@@ -0,0 +1,27 @@
+# Ink wash and dry brush
+
+Paper, expressive ink, bleed and meaningful unpainted space.
+
+## Material and construction
+
+Begin with paper, then add ink; do not fill the entire image before adding a paper filter. Sample strokes by arc length with a tapered brush tip, bristle gaps and dry-brush white breaks. Add a softly blurred ink-bleed layer locally. Use restrained seals or calligraphy only when appropriate, with meaningful text composed separately.
+
+## Palette
+
+Warm paper, black and diluted gray ink, with an optional small vermilion seal.
+
+## Motion
+
+Write a stroke into existence, let ink spread after contact, and animate a leaf, shrimp antenna or rippling water. Keep dry-brush gaps attached to the stroke instead of flickering independently.
+
+## Transition
+
+A brush stroke writes the next frame or a controlled ink wash spreads.
+
+## Failure cases and adaptation
+
+Do not import an oil-paint layer order. Large negative space is part of the medium. The source figures were less expressive than its brushwork: prepare a convincing subject still and apply the same ink edge treatment to it.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/17_ink.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/klimt-gold.md

@@ -0,0 +1,27 @@
+# Decorative gold and geometric ornament
+
+Gold leaf, spirals, patterned robes and selective naturalistic features.
+
+## Material and construction
+
+Combine polished gold texture with spiral and geometric pattern fields. Keep faces and hands more legible than the ornament. Draw patterns into an offscreen layer, clip once to the costume silhouette, then composite it back. Repeated source-in operations on individual motifs can erase earlier work.
+
+## Palette
+
+Gold, warm cream, black and restrained jewel accents; vary metallic value rather than scattering unrelated hues.
+
+## Motion
+
+Ornamental eyes blink, small leaves drift, a gold sweep crosses the robe and spiral ornament unfolds. Clothing follows the body anchors, with highlights clipped to the gold material.
+
+## Transition
+
+A spiral expands or gold ornament grows into the next frame.
+
+## Failure cases and adaptation
+
+Pattern clipping and hand attachment were weak in the source study. Ornament should not obscure faces. Do not wash paper color over a subject with source-atop; mask the texture and use the intended multiply blend instead.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/18_klimt.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/matisse-cut-paper.md

@@ -0,0 +1,27 @@
+# Gouache cut paper
+
+Scissor edges, saturated sheets, pinned shadows and organic cut shapes.
+
+## Material and construction
+
+Sample contours every 14–24 px with about ±2–2.5 px seeded edge variation, then join with straight scissor segments. Build seaweed from a tapered stem and alternating round capsule lobes, unioned before fill and shadow. Add a short paper shadow around (5,6) px with blur 4 and .3 alpha. Keep gouache brush texture restrained.
+
+## Palette
+
+Cobalt #1D4FB8 / #4F95DC, green #1F9A5A, yellow, black, white and a small red accent.
+
+## Motion
+
+Pinned seaweed rotates around its base, paper edges curl, a bird wing flips and a loose scrap turns as it falls. ScaleX can approach zero during a flip, but preserve a coherent path and avoid the subject face.
+
+## Transition
+
+A colored sheet tiles or sweeps over the next composition.
+
+## Failure cases and adaptation
+
+Width-modulating a spine produced caterpillars rather than lobed seaweed in the source. Heavy long brush marks can overwhelm the clean cut-paper treatment. Place type on an actual contrasting paper piece when the background is saturated.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/30_matisse.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/monet-water-garden.md

@@ -0,0 +1,27 @@
+# Water-garden impressionism
+
+Water reflections, lilies, willow shadows and color-built light.
+
+## Material and construction
+
+Paint water in blue/lavender bands, then horizontal short strokes without outlines. Source water marks used cell 9, length 34 and width 6.5 px. Treat bridge, window and subject in separate finer passes so structure survives the foliage. A dark green backing behind the bridge and lighter green structural lines keep it visible. Use warm light and violet shade on characters.
+
+## Palette
+
+Water #5B6FB5 / #7D8FD0 / #9DB5E0 / #6AA0A8 / #B7A6D8, green leaf groups and pale pink/white flowers.
+
+## Motion
+
+Short horizontal light marks shimmer, lilies drift slowly, willow reflections wave and elliptical ripples expand. Wisteria should hang as clustered dots rather than smooth capsules. Keep floor and architecture static while water responds.
+
+## Transition
+
+A water ripple or reflection dissolve connects scenes.
+
+## Failure cases and adaptation
+
+Whole-screen 6 fps boiling became noisy in the source. Localize movement to water and foliage. Mixing a pond mural and a separate room can confuse space; choose a coherent composition for the new subject. Give text its own quiet backing.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/28_monet.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/pixel-art.md

@@ -0,0 +1,27 @@
+# 8-bit pixel art
+
+Low-resolution silhouettes, limited colors and discrete sprite acting.
+
+## Material and construction
+
+Draw at a small logical resolution such as 240×135 for 1920×1080 output; enlarge with nearest-neighbor sampling. Use a limited palette around 14 colors and a one-logical-pixel outline per part. Place features on the pixel grid. Reuse supplied or generated sprite sheets with their actual frame rectangles and timing.
+
+## Palette
+
+A small indexed palette with deliberate shared ramps; avoid independently quantizing each frame.
+
+## Motion
+
+Step a tail through a few poses, bounce a collectible or question block, and animate a concise HUD event. Use discrete poses rather than fractional subpixel drift.
+
+## Transition
+
+Reveal pixel rows, assemble tiles or make a game-like screen change.
+
+## Failure cases and adaptation
+
+Smooth scaling and antialiasing blur the pixel grid. Overlapping outlined pieces need explicit occlusion order. Essential text needs a compatible bitmap or deliberately rasterized font, and sufficient logical pixels per character.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/14_8bit.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/polka-dot-infinity.md

@@ -0,0 +1,27 @@
+# Polka dots and infinity space
+
+Hand-irregular dot fields, repeating depth and ribbed organic forms.
+
+## Material and construction
+
+Use seeded irregular hexagonal grids with deliberate large and small dots. Squash floor spacing in perspective; build infinity lights as a coherent multi-depth grid rather than random bokeh. For a pumpkin-like form, align dot size and spacing to the ribs, largest near the central visible face.
+
+## Palette
+
+High-contrast dot/ground pairs, plus a restrained multi-color light grid when the scene calls for infinity-space depth.
+
+## Motion
+
+A radial wave can scale dots by .82+.28 sin(.012 distance−6t). Let a ribbed form breathe about its base and send a traveling wave through the dot field. Use the true two-row hex-grid period for scrolling.
+
+## Transition
+
+Dots grow to reveal the next image through a clipped mask.
+
+## Failure cases and adaptation
+
+Limit a transition dot radius relative to its spacing; around .47 of the gap avoids an uncontrolled whiteout in the source setup. Preserve subject/background contrast when both carry dots. Random bokeh alone loses the organized spatial idea.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/21_kusama.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/pop-print.md

@@ -0,0 +1,27 @@
+# Pop-art print
+
+Ben-Day dots, heavy ink contours, flat color and bold comic emphasis.
+
+## Material and construction
+
+Build repeatable dot textures as small pattern tiles. A 45-degree grid can use corner and center dots; example pitches are 24 px for a wall and 13 px for skin at 1080p. Cache the background. Draw flat pieces with roughly 5 px dark contours, white highlight slashes and clipped diagonal hatching. Treat ink and color registration deliberately.
+
+## Palette
+
+Ink #141414, paper #F3EEDB, blue #2F5CC8, yellow #F7D52C, red #D8232A; optional quadrant colors for a screen-print sequence.
+
+## Motion
+
+Cycle colors in a four-panel print, stagger thought-bubble dots, quantize an onomatopoeia burst and add one clear character anticipation. Keep the burst away from the face and essential evidence.
+
+## Transition
+
+A lightning-shaped halftone reveal or comic-panel cut.
+
+## Failure cases and adaptation
+
+A large yellow hair mass reads as cloth without distinct curls and internal C-shaped ink arcs. When a body crouches, move its head anchor with it. Screen-space and object-space pattern alignment are choices: avoid unintended texture swimming.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/13_pop.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/post-impressionist-flow.md

@@ -0,0 +1,27 @@
+# Post-impressionist flow strokes
+
+Directional impasto and expressive color fields, including a Van Gogh-inspired treatment.
+
+## Material and construction
+
+Use a flat composition as a color source, then repaint in several passes with region-specific vector fields: vertical wall marks, perspective floor marks and star vortices. Give each region its own multi-color swatch. Dark stroke rims and small bright ridges create paint body. Process props and characters separately with finer marks, then restore visible features.
+
+## Palette
+
+Contrasting cobalt/ultramarine, warm yellow and ochre, with distinct regional color groups; the palette follows the chosen artwork or subject.
+
+## Motion
+
+Stars rotate through vortices, a lamp halo expands, vegetation sways and the main subject acts. Keep stroke locations stable while the flow field changes continuously. Cache static work and profile the full multi-pass renderer.
+
+## Transition
+
+A directional swirl carries paint into the next image.
+
+## Failure cases and adaptation
+
+Single-color jitter produces colored noise rather than expressive painting. Large brushes erase small props. Full-frame seed changes flash. The source implementation was among the expensive styles at roughly 110 ms/frame; that is historical context, not a current AnimaGuru benchmark.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/09_postimp.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/radiant-line-figures.md

@@ -0,0 +1,27 @@
+# Radiant line figures
+
+Bold equal-weight contours, flat colors, dancing silhouettes and radiating action marks.
+
+## Material and construction
+
+Use roughly 13 px rounded black contours at 1080p, flat fills and a simple ground. Union capsule limbs before outlining to avoid holes at joints. Keep internal detail sparse and the contour weight consistent across figures, props and action marks.
+
+## Palette
+
+Yellow #FFD51C, green #16A54A, black and clear red/blue/pink accents.
+
+## Motion
+
+Hard-cut among four dance poses at about 8 fps rather than interpolating them smoothly. Add short radiating marks with stepped angles. Let objects squash around contact points on selected beats, with slight phase differences; a heart or small crawling figure can provide a secondary loop.
+
+## Transition
+
+A bold figure silhouette or radiating shape becomes a portal.
+
+## Failure cases and adaptation
+
+Fine hair lines contradict equal-weight drawing and may cross a moving face. Rounded limb endcaps must join without gaps. The source figure proportions were not fully characteristic, so test silhouette proportions before adding more activity.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/31_haring.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/rembrandt-chiaroscuro.md

@@ -0,0 +1,27 @@
+# Chiaroscuro and impasto light
+
+Figures emerge from warm darkness through selective window and candle illumination.
+
+## Material and construction
+
+Paint a fully readable lit base with broad background strokes, finer props and modeled subjects. Multiply it by a lighting map: low ambient light plus broad window, face, chest, table and candle pools. Add subtle separation light behind the dark side of the subject. Soften lost edges with a blur/clear pair, preserving facial features. Place bristle impasto highlights only in lit regions.
+
+## Palette
+
+Deep umber ambient light, warm golden highlights and restrained red/cream costume notes.
+
+## Motion
+
+Candle flame leans and changes height with seeded continuous noise; light pools shift slightly, clouds dim the window, and dust appears only within the shaft. Let the subject perform one small action in the pool.
+
+## Transition
+
+An iris of light or darkness reveals another composition.
+
+## Failure cases and adaptation
+
+Three neat highlight stripes resemble stickers; use many irregular fine bristles and broken ridges. Recopying a rectangular crop restores unwanted unpainted background: keep props on separate layers. Dense grain dirties faces. Check hand → cup → thumb occlusion order.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/32_rembrandt.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/renaissance-sfumato.md

@@ -0,0 +1,27 @@
+# Renaissance sfumato
+
+Soft edges, modeled volume, aerial perspective and glazed oil texture.
+
+## Material and construction
+
+Paint perspective architecture and a cool receding landscape, then cache a lightly softened room. Render characters into a separate layer: a soft pass plus a clear pass preserves faces. The upstream revision reduced character blur from 2.6 to 1.6 px and raised the clear pass from .38 to .82 after the subjects became less legible than the distant view. Add restrained canvas, crackle, warm glaze and vignette.
+
+## Palette
+
+Umber and muted red indoors, blue-gray distant hills, a warm window-light pool; keep floor contrast subordinate to the subject.
+
+## Motion
+
+A ribbed ornithopter flaps outside, water catches light, birds cross and dust moves through a cached soft light shaft. Wing spars and radiating ribs communicate the mechanism better than a paper-plane silhouette.
+
+## Transition
+
+Travel through the arched window or use a soft light reveal.
+
+## Failure cases and adaptation
+
+Dense directional strokes drift toward post-impressionism. Over-dark inner contours make pale bodies look glassy. Cache blurred beams rather than painting hard light polygons each frame. Atmospheric softness should preserve the focal plane.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/06_renaissance.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/roman-mosaic.md

@@ -0,0 +1,27 @@
+# Roman mosaic
+
+Contour-following tesserae, stone variation and grout. Useful for architectural or historical art worlds.
+
+## Material and construction
+
+Compute tile layout once. Follow distance-field contours around important shapes; use a weighted L-infinity Voronoi construction for less mechanical stone boundaries. Tile half-width can start 8–28% above half the nominal spacing so bisectors carve convincing grout. Animate the sampled color or subject beneath the fixed stone layout, rather than changing tile topology every frame.
+
+## Palette
+
+Warm limestone, terracotta, cream and dark stone; use regional stone palettes instead of uniform color noise.
+
+## Motion
+
+Let a volcanic plume, sparks or water move through the tesserae. A border conveyor can loop, provided its baked pattern period matches the actual tile period. Sub-tile color travel may disappear through quantization; moving the tiled border itself makes travel visible. Preserve small facial features with a finer local treatment that does not form a rectangular patch.
+
+## Transition
+
+Flip or replace tiles in a coherent wave, keeping grout visible through the change.
+
+## Failure cases and adaptation
+
+A square grid with random colors reads as pixels. Rebuilding the layout causes boiling. Thin lettering breaks: use broad monoline forms with strokes at least two tesserae wide. Props narrower than roughly three tiles can merge into the scene; the upstream revision used a separate finer-tiled prop inset moving with the hand. Enlarge the dynamic redraw region beyond every affected tile, and keep the same wall layout and tile scale inside and outside that region. Otherwise its boundary becomes a visible patch.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/04_roman.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/rubber-hose.md

@@ -0,0 +1,27 @@
+# Rubber-hose cartoon
+
+Pie-cut eyes, noodle limbs, white gloves, contact squash and old-film cadence.
+
+## Material and construction
+
+Cache a simple vintage setting. Use round faces, constant-width curved limbs and simplified gloves. Render the complete scene, then desaturate and apply a warm faded-paper tint consistently. Quantize character time to 12 fps while film wear can update at 24 fps.
+
+## Palette
+
+A controlled monochrome range with warm paper #F2E8D2; keep white gloves and props in the same treatment.
+
+## Motion
+
+Bounce around each object’s contact point on the chosen beat. Source squash used about 9% compression with inverse horizontal expansion and slight airborne stretch. Offset phases for secondary objects. Animate an actual pose, dancing flower or smoke ring before adding gate weave, scratches and dust.
+
+## Transition
+
+A rounded film-gate iris or a cartoon smoke transition.
+
+## Failure cases and adaptation
+
+A high motion percentage can come almost entirely from film noise; inspect clean character movement separately. Smooth angular body outlines where the cartoon needs curves. Apply monochrome treatment after all subjects, or individual props retain inconsistent color.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/33_rubberhose.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/seurat-pointillism.md

@@ -0,0 +1,27 @@
+# Pointillist optical mixing
+
+Equal-size discrete pigment dots whose juxtaposition forms color.
+
+## Material and construction
+
+Lay a softly filtered flat base under a fixed hexagonal dot grid. Source pitch 10, row spacing 8.66 and radius 4.5 px produced about 24,000 dots. Select from a pure-color palette including pale mixtures. Project the target color onto candidate palette-pair segments, clamp the mixing weight to .2–.8, then choose a pigment by seeded hash. Cache quantized target-color matches and batch one path per pigment.
+
+## Palette
+
+Saturated blues, greens, yellow, orange, red-violet and purple, plus pale yellow, peach, green, blue and pink. Pure dark/saturated colors alone make light regions snowy.
+
+## Motion
+
+Let boats drift, a butterfly flap and river dots catch light. A restrained subset of dots may switch paired colors at stepped times; source 6 fps/7% was more visible than 4 fps/3%. Preserve the overall calm.
+
+## Transition
+
+Dots crystallize into the next image.
+
+## Failure cases and adaptation
+
+Hard palette-pair boundaries form bands: add small seeded target-color dithering before matching. An optional 8% complementary-dot pass and edge light transmission enrich the surface. Equal pure-color dots distinguish this from impressionist dabs; a regular computer grid may still need subtle fixed jitter.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/29_seurat.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/shadow-puppet.md

@@ -0,0 +1,27 @@
+# Chinese shadow-puppet theatre
+
+Translucent carved leather, riveted joints, visible rods and warm cloth backlight.
+
+## Material and construction
+
+Create a cached cloth screen and prop layer. For each leather piece: fill, carve with destination-out inside a clip, restore a protective leather rim, then outline in deep brown around 2.6 px. Stamp a soft offset shadow followed by the multiply leather layer. Add rivets and rods. Keep carving on isolated offscreens so it does not cut holes in the final frame.
+
+## Palette
+
+Red #C0352A, green #2F7D4C, amber #D99A2E, yellow #E2BD4A and blue #2C5A8A against warm backlight.
+
+## Motion
+
+Step the main action near 10 fps. Rotate the head around its neck rivet, let a loose wrist and cup swing together, and delay each segment of a multi-jointed tail. A rod-operated butterfly flips its wings; lamp variation moves the cloth light pool.
+
+## Transition
+
+A backlight shift or articulated leather screen sweep.
+
+## Failure cases and adaptation
+
+White fur can be cut out, retaining only the leather edge. Multiply overlaps legitimately darken, but keep near-black overlaps away from the focal face or costume. Rods should travel down/out rather than cross other characters. Many offscreen multiply passes were expensive in the source.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/34_shadowpuppet.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/surrealist-soft-forms.md

@@ -0,0 +1,27 @@
+# Surrealist soft forms
+
+Hard light, long shadows and soft objects that violate their expected rigidity.
+
+## Material and construction
+
+Cache the setting; layer an animated shadow, props, warped object and subject. Render a long shadow into one mask before applying alpha/blur, avoiding dark overlaps between separate pieces. Deform every part of a soft clock through the same mapping: rim, dial, hands and indices must agree.
+
+## Palette
+
+Sky #22407C / #5F88BD / #E9DCAE, sea #8AA6B8 / #4F6F86, ground #D2AD70 / #7A5129 and aged gold.
+
+## Motion
+
+Let a clock visibly sag, ants travel a coherent path or a stilt-legged animal step with phased legs. A source clock used top compression .28–.32 and a pronounced sag over about 1.2 s. Round the droplet profile outward rather than pulling sharp cones down.
+
+## Transition
+
+Rounded liquid drops bring in the new image.
+
+## Failure cases and adaptation
+
+Warping only the outline leaves the clock hands rigid and breaks the illusion. Maintain ground contact and a consistent light vector. Generic flat vector subjects can weaken the painted world; prepare them in the same medium.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/23_dali.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/vaporwave-vhs.md

@@ -0,0 +1,27 @@
+# Vaporwave and VHS neon
+
+Deliberate retro neon, perspective grids, banded sunset and analog-video distortion.
+
+## Material and construction
+
+Compose the scene offscreen, then add channel separation and tracking as a post layer. Neon tubes combine a broad blurred glow with a thin near-white core. Cut widening horizontal gaps from a gradient sunset. Use a projected grid with a stable vanishing point and distance-dependent line spacing; animate phase modulo the true repeat period.
+
+## Palette
+
+Magenta #FF4FD8, hot pink #FF2F8F, cyan #3FF6FF, deep purple #12052A / #2A0B4A, yellow #FFD84A and orange #FF8A3A. This palette is intentional for this treatment.
+
+## Motion
+
+Scroll the grid, drift sunset bands, flicker selected sign tubes and pass a bounded VHS tracking band. RGB offsets can start near 3 px with occasional larger glitches. Keep exact UI labels composed above destructive post effects when needed.
+
+## Transition
+
+A VHS tracking tear or signal interruption changes the shot.
+
+## Failure cases and adaptation
+
+Check glyph coverage for arrows and non-Latin neon signs. A garment too close to skin color can become unreadable. Outlining every component produces internal neon seams; derive an outer silhouette when that is the intended effect. Full-frame noise is not evidence of meaningful animation.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/26_vaporwave.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).

package/skills/animaguru-art-motion/references/art/watercolor-cel-world.md

@@ -0,0 +1,27 @@
+# Watercolor background with cel characters
+
+Layered watercolor environments, warm linework and observed natural motion.
+
+## Material and construction
+
+Displace polygon edges in several seeded midpoint passes, reducing amplitude by about .6 each pass. Layer transparent washes with tide-line edges. Example walls use five washes near .4 alpha and broad soft blooms; wood uses fewer tighter washes. Paint white highlights source-over rather than multiply. Cel subjects get a simple base, crescent shadow and warm brown contour.
+
+## Palette
+
+Natural greens, warm earth, translucent pale washes and clean light accents; avoid a uniformly moldy sepia cast.
+
+## Motion
+
+A gust bends the lower curtain edge, grass receives a traveling phase, clouds drift, and dappled light, hair and leaves share the same breeze. Cache static watercolor work and keep its seeds fixed.
+
+## Transition
+
+A wet-paper bloom or foreground foliage reveals the next scene.
+
+## Failure cases and adaptation
+
+Hard rectangular shadow seams and flat vector faces break the painted setting. Soften only what needs it. Apply compatible edge/material treatment to characters. The upstream background was stronger than its figures; do not treat that demo as finished character-quality evidence.
+
+Parameters above are upstream study values at approximately 1920×1080 unless stated otherwise. Scale them to the delivery format and check the actual result. These are English production adaptations, not bundled renderers or claims of validated AnimaGuru implementations.
+
+Adapted from [the upstream card](https://github.com/alchaincyf/huashu-art-motion/blob/2571ee0a9a0467c4d99e935ea9e0fadda19210b0/references/%E9%A3%8E%E6%A0%BC%E9%85%8D%E6%96%B9/25_ghibli.md) under the [MIT license](../../LICENSE). See [provenance](../sources.md).