@docyrus/animaguru 0.6.0

package/skills/animaguru-cartoon-character-animation/SKILL.md

@@ -0,0 +1,35 @@
+---
+name: animaguru-cartoon-character-animation
+description: Create animated films with expressive illustrated characters, readable poses and coordinated props. Use for the Cartoon Character Animation visual style in an AnimaGuru video workspace.
+---
+
+# Cartoon Character Animation
+
+Use illustrated characters and clear acting to express the user's message. Casting, relationships, mood and narrative come from the Explainer, Advertisement or General Video brief; a cartoon treatment does not imply a child audience or comedy.
+
+## Establish the direction
+
+Follow the sibling `animaguru-video` category → style → quick interview flow for new films. Reuse supplied choices and briefs; honor an explicit request to skip the interview. Share [the style preview](https://animaguru.com/animation-styles#cartoon-character-animation). Add missing visual choices to the shared interview:
+
+- Which characters, objects or mascots are needed, and are approved designs available?
+- What line weight, palette, shading and illustration references should guide the cast?
+- Should acting be subtle, broad or stylized? Is speech shown on camera, off screen, or absent?
+
+## Design and build
+
+- Read `animaguru` for generation, reference imports, current model discovery and custom motion jobs. Use `hd-cartoon` where appropriate. Record a concise model sheet: silhouette, proportions, clothing, palette, facial features and view. Reuse approved assets to preserve identity.
+- Plan performances as readable poses with anticipation, action and a resolved hold. Choose the necessary motion from the actual story. AnimaGuru's built-in game actions are not a complete acting vocabulary; use supported custom motions for pointing, considering, presenting and other brief-specific actions.
+- Read `pixijs`, sprite and container guidance. Choose spritesheets, separate articulated parts or generated video based on the performance. Simple limb transforms are precise; nuanced facial acting may need generated frames or video. Do not apply rubbery stretching to every design.
+- Keep character anchors, apparent height, facing and prop ownership consistent. Separate rear limbs, torso, front limbs and held props where needed. Track contact points to prevent floating feet and sliding hands.
+- Derive spritesheet frame indices from absolute time, respecting per-animation frame counts, holds and loop behavior. Match the performance cadence to the selected design while keeping the film export fps explicit.
+- Give acting priority over background motion. Aim eye lines and gestures toward the intended focus. Use reaction holds rather than constant idle movement for every character.
+- For dialogue, use verified phoneme/viseme or performance timing when lip sync is required. A talking loop is not accurate lip sync. If only off-screen narration is needed, keep mouth movement appropriate to the scene.
+- Keep typography and captions editable, and register characters, props and scene layers in the resource timeline. Generated video must obey the shared asynchronous frame-decoding contract.
+
+## Review and completion
+
+Follow `animaguru-video` for production records, audio and exports. Inspect character identity, hands, eye lines, contact points, occlusion, pose readability and loop seams. Check movements at full speed and selected extremes. Deliver approved cast assets, coherent performances and the working HTML5 film; export MP4 when requested.
+
+## Additional visual treatments
+
+Read [the material and motion variants](references/variants.md) when the brief calls for a specialized treatment. Choose the relevant recipe without imposing its example story.

package/skills/animaguru-cartoon-character-animation/references/variants.md

@@ -0,0 +1,17 @@
+# Character finish and cadence
+
+Choose one coherent treatment from the brief; these are visual/motion recipes, not mandatory stories. They apply to Explainer, Advertisement and General Video.
+
+## Rubber-hose movement
+
+Continuous noodle-like limbs and elastic action can make objects expressive. Sample drawing poses deliberately while the camera remains smooth. Keep stroke weight stable in screen space and preserve a clear silhouette; no existing character shapes or mandatory jazz/comedy sequence.
+
+## Mid-century limited animation
+
+Use separate off-register fill/line plates and held bodies with active hands, eyes and mouth. Keep hand contact with props and build anticipation/action/settle from poses. Classroom narration and instructional scene order are choices from the brief.
+
+## Cel and dialogue performance
+
+Define pose, mouth and expression sets with explicit cadence, plus prepared painted backgrounds. A held cel can still act through eyes or a carefully timed gesture. Lip synchronization requires recorded speech timing or verified mouth cues; a generic talk loop is not lip sync.
+
+Use the bundled absolute-time helpers in `src/studies/timing.ts` where appropriate. Keep actual resources and cues in the timeline, then inspect cold seeks, full-speed motion and exported frames. The shared video skill governs interviews, generation scope and delivery.

package/skills/animaguru-cutout-animation/SKILL.md

@@ -0,0 +1,33 @@
+---
+name: animaguru-cutout-animation
+description: Create layered cutout animation using photographic, illustrated, or paper-style assets, collage composition, and deliberate 2D motion. Use for the Cutout Animation visual template with any audience, subject, or communication goal.
+---
+
+# Cutout Animation
+
+This template defines the visual treatment and animation language: separated silhouettes, layered composition, material edges, depth, and purposeful movement. It works with Explainer, Advertisement and General Video projects. The subject, audience, tone, message, narrative structure, duration, language, and call to action come from the user's brief.
+
+## Establish the brief
+
+For a new film, **follow the category-then-style selection flow and quick interview in the sibling `animaguru-video` skill before production**. This skill being installed does not mean the user has selected it. Share the [animation-style gallery](https://animaguru.com/animation-styles#cutout-animation) when the user wants to see this style. Reuse supplied choices and completed briefs; honor an explicit request to skip the interview. Include the visual choices below in the brief interview when they are still open:
+
+- Cutout material: photography, illustration, printed paper, mixed-media collage, or a supplied reference.
+- Composition: graphic and flat, or layered with visible depth, shadows and parallax.
+- Motion: smooth editorial movement, articulated cutout animation, or stepped stop-motion cadence.
+- Visual constraints: brand palette, typography, texture, edge finish, characters or objects, and required asset references.
+
+A compact default direction can combine clean silhouettes, overlapping planes and eased 2D movement. Derive casting, mood, colors, tempo and scene design from the brief. Do not infer an audience, relationship between characters, story arc, fixed scene count, narration style or CTA from this template.
+
+## Apply the style
+
+Read [visual and motion guidance](references/visual-style.md) when designing scenes. Read [the production recipe](references/production-recipe.md) when generating and integrating cutout assets. Follow the sibling `animaguru-video` skill for the interview, source research, production records, resource timeline, audio and export workflow.
+
+Use image tools for bitmap artwork, cutouts and textures, and PixiJS for composition, typography, masks and motion. Treat people, products, objects, landscapes and abstract forms as equally valid cutout subjects. Choose coherent edge treatment, material, lighting and movement so the film feels intentionally assembled.
+
+Use the user's selected generation models and budget. Otherwise inspect the current AnimaGuru catalog and configured providers to select models suited to the assets and motion. MiniMax H3 Max (`fal/minimax-h3-max`) is one option for animated photographic assets when available; cutout motion can also be built directly in PixiJS from still layers. Validate supported settings before submission.
+
+Deliver the film described in the brief with consistent cutout visuals and intentional animation. Keep story development responsive to the project's purpose rather than borrowing the narrative of an example film.
+
+## Additional visual treatments
+
+Read [the material and motion variants](references/variants.md) when the brief calls for a specialized treatment. Choose the relevant recipe without imposing its example story.

package/skills/animaguru-cutout-animation/references/production-recipe.md

@@ -0,0 +1,37 @@
+# Cutout production recipe
+
+## Plan reusable layers
+
+Start from the agreed brief and scene plan. List the required subjects, moving parts, props, background planes and textures. Decide which movements need generated video and which can use deterministic Pixi transforms. Reuse assets where appropriate without forcing a fixed cast, number of clips, duration or narrative pattern.
+
+Read the installed AnimaGuru skill and inspect the current catalog, command schema, chosen project and asset list. Use the requested project and models. Save generation requests and receipts, then keep returned IDs, model choices, local files and hashes in the asset manifest. Reconcile an uncertain submission before retrying paid work.
+
+## Generate cutout artwork
+
+Specify the asset's material, framing, silhouette, pose or orientation, perspective, lighting, edge treatment and transparent/keyed background. Match these choices across assets. For articulated motion, plan separately drawable parts and sufficient overlap at joints before generating them.
+
+An object prompt structure to adapt:
+
+> The specified object, isolated with its complete silhouette and generous margin, viewed from the chosen angle, using the project's selected photographic or illustrated treatment. Consistent lighting and perspective with the other assets. Clean transparent background, no text, no incidental scene elements.
+
+For people or animals, add only the casting, expression, wardrobe and pose needed by the brief. For paper or illustrated work, describe material and edge character. Avoid cropping parts that the planned movement will reveal.
+
+Transparent source images are useful. For a video model that flattens alpha, use a uniform key color absent from the subject and props. Specify a locked camera and stable scale when the clip will be composited into an editorial scene. Inspect the output rather than assuming a prompt guarantees a clean key.
+
+## Animate and composite
+
+Keep a generated clip's motion request focused: describe the movement, its starting and ending pose, and whether a clean loop is needed. Choose duration and resolution from the selected model's supported profile. Use returned asset references rather than guessed identifiers.
+
+For still layers, group related pieces in Pixi Containers and define pivots, masks and depth order explicitly. Compute translation, rotation, scale and stepped timing from absolute film time. Use seeded variation only where it serves the chosen material treatment.
+
+Use source video when it preserves useful performance detail; use sprite sheets or still layers where they suit the motion. For green/magenta keying, read `pixijs-custom-rendering` and `pixijs-assets`. Apply a soft matte and conservative spill suppression; inspect fine edges and preserve genuine subject colors. Set the key explicitly per asset.
+
+Exact-frame capture must pause video, seek to the required clip time, await decoding with a bounded timeout, update the texture source, and then render. Clamp or loop within the actual clip duration. Check a loop seam before repeating it.
+
+## Integrate and verify
+
+Keep assets local in `public/assets/` and resolve them with `publicAsset()` so HTML5 embeds work in nested hosting folders. Register each resource and its real cue times in `src/film-resources.ts`.
+
+Follow the shared video skill for narration, music, factual claims and export delivery when those are part of the brief. Cutout Animation does not prescribe a voice, soundtrack, message, story structure or ending.
+
+Replace the starter's demo scenes, update film timing, preserve the asynchronous capture API, and inspect representative full-resolution frames at transitions and movement extremes. Review the final film in the intended format and produce the requested HTML5 or MP4 deliverable through the supplied export pipeline.

package/skills/animaguru-cutout-animation/references/variants.md

@@ -0,0 +1,17 @@
+# Paper mechanisms and graphic scenery
+
+Choose one coherent treatment from the brief; these are visual/motion recipes, not mandatory stories. They apply to Explainer, Advertisement and General Video.
+
+## Fold, cut and unfold
+
+Prepare the fold axis and back face. A mirrored pattern multiplies through a hinge/mesh transformation rather than merely sliding duplicate cards. Preserve holes, scissor edges and paper thickness during the transition.
+
+## Illuminated paper depth
+
+Give each layer a prepared alpha silhouette, depth and light response. Layered tint/transmission can suggest a lightbox in 2.5D. Full physical lamps, free camera orbit and real cast shadows remain deferred.
+
+## Typographic scenery
+
+Use letters or words as measured scene structures, with original glyph geometry and stable cut edges in local coordinates. Puppets may move in held drawings while the camera moves continuously. Spy chases, festivals, mandatory red and children's stories are optional subjects.
+
+Use the bundled absolute-time helpers in `src/studies/timing.ts` where appropriate. Keep actual resources and cues in the timeline, then inspect cold seeks, full-speed motion and exported frames. The shared video skill governs interviews, generation scope and delivery.

package/skills/animaguru-cutout-animation/references/visual-style.md

@@ -0,0 +1,34 @@
+# Cutout visual and motion language
+
+Use these techniques to realize the user's subject and message. Select a coherent subset for the film; its narrative order and pacing come from the brief.
+
+## Material and silhouette
+
+- Photography gives subjects recognizable surface detail; illustration permits simplified forms; printed paper adds a tactile finish. Combine them when the contrast is deliberate.
+- Choose crisp, torn, outlined or visibly handmade edges. Keep edge quality consistent across assets and inspect fine detail at the intended viewing size.
+- Separate foreground subjects, props and background planes. Leave sufficient margins for movement and use clean transparency or a controlled key color.
+- Match lighting, perspective and scale across photographic assets. Derive colors, type and textures from the project's identity and visual references.
+- Apply grain and paper texture selectively so they support the material without obscuring subjects or type.
+
+## Composition and depth
+
+- Use overlap, restrained shadows and differing layer scales to establish depth. Keep contact shadows anchored to the relevant surface.
+- Frame people, objects, products, diagrams or abstract shapes according to the scene's purpose. A human cast is optional.
+- Use negative space and clear visual hierarchy. Keep important imagery and type inside the output format's safe area, including caption space when captions are part of the brief.
+- Choose layouts for the actual aspect ratio. Recompose for portrait or square rather than compressing a landscape composition.
+- Keep essential type in PixiJS so it stays editable, legible and localizable. Generated artwork should not carry critical wording.
+
+## Movement
+
+- **Editorial motion:** use eased translation, rotation and scale, with restrained overshoot or settling when appropriate.
+- **Articulated cutouts:** separate moving parts, place pivots at plausible joints or hinges, and layer overlaps to conceal seams. Preserve silhouette and material consistency during movement.
+- **Stepped motion:** sample transforms at a deliberate cadence for a stop-motion feel. Keep the main timeline deterministic and choose the cadence to suit the visual reference.
+- **Parallax:** move foreground and background planes at different rates. Keep camera movement controlled and relevant to what the viewer should notice.
+- **Asset performance:** use generated video when nuanced gestures or organic motion are needed. Animate still layers in Pixi when transforms, reveals or articulation can achieve the result more precisely.
+- **Transitions:** slide, layer, peel, reveal, occlude or match shapes in ways that suit the material. Use a consistent transition vocabulary.
+
+Give movements clear beginnings, readable holds and resolved endings. Choose energy and rhythm from the brief; the cutout style supports quiet, playful, technical, dramatic and many other treatments. Avoid perpetual motion that competes with the intended focus.
+
+## Visual review
+
+Inspect silhouette edges, key spill, joints, shadows, overlaps, texture scale, type fit and repeated-motion seams. Check transitions and the first/last frames at full resolution. Preview the work in its intended aspect ratio and ensure movement continues to communicate when viewed at a smaller size.

package/skills/animaguru-dark-tech-stage/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: animaguru-dark-tech-stage
+description: Build Dark Tech Stage visuals and motion in an AnimaGuru PixiJS video workspace. Use when this visual style is chosen for an explainer, advertisement or general video.
+---
+
+# Dark Tech Stage
+
+Dark presentation stages, crisp components and controlled accent light with precise assembly and focused reveals. Subject, audience, emotional tone and story come from the user's brief. This style supports Explainer, Advertisement and General Video.
+
+## Establish direction
+
+Follow sibling `animaguru-video` for purpose → style → compact interview before new production. Reuse explicit choices and completed briefs; honor a requested interview skip. Share [the style gallery](https://animaguru.com/animation-styles#dark-tech-stage). Fold missing visual questions into the shared interview:
+
+- Which brand colors and type should govern the stage?
+- Should the subject use real interface states, objects, or abstract systems?
+
+## Build the look and movement
+
+Read `pixijs` and the relevant Graphics, Text, Container, Assets or Mesh skills. Read `animaguru` before generating assets; discover current models and estimates, respect the user's budget, and reuse completed media. Essential text and exact numerical geometry belong in authored canvas content.
+
+- Give each subject a presentation plane and a quiet text region; accent light marks hierarchy rather than decorating every edge.
+- Derive accumulated items and their final grid positions from one table; arrivals can land on a cue while the camera remains continuous.
+- Preserve real interface labels and supported states. Abstract panels are conceptual illustrations, not evidence of shipped features.
+
+Use the [production recipe](references/recipe.md) for the bundled original study and port-specific decisions. It is a technique example, not a prescribed film. Keep motions a function of absolute seconds; separate drawing cadence from camera/light cadence when the material calls for it. Prepare assets locally, preserve async seeking and register actual resource/cue times. Keep essential type editable and localized.
+
+## Review and delivery
+
+Inspect low-contrast panels, uncontrolled glow and text masked by foreground cards; random-order seeks before and after grid assembly. Check representative full-resolution frames, moving playback, desktop/mobile fit and cold seeks. Follow `animaguru-video` for production records, audio mixing and HTML5 delivery; export MP4 when requested. Describe the actual 2D/2.5D rendering method accurately. Physically rendered 3D remains deferred.

package/skills/animaguru-dark-tech-stage/references/recipe.md

@@ -0,0 +1,23 @@
+# Dark Tech Stage production recipe
+
+The workspace includes an executable original study in `src/studies/scenes.ts`, selected with `dark-tech-stage`. Its drawing commands are consumed by `src/studies/pixi-renderer.ts`; geometry and time sampling are shared with the gallery renderer. All copy lives in `shared/locales/en/style-studies.json`.
+
+In an untouched starter, run `npm run example -- dark-tech-stage`, then `npm run dev`. The command refuses to replace an authored film, retains starter backups in `production/example-backup/`, and writes an 8-second study with its own resource tracks. `npm run export:html` and `npm run export:mp4` use this installed film normally. To use the recipe inside an existing film, import `studyFrame()` and `StudyRenderer`, then compose the selected scene into the project's timeline instead of running the installer.
+
+## Adaptation decisions
+
+- Give each subject a presentation plane and a quiet text region; accent light marks hierarchy rather than decorating every edge.
+- Derive accumulated items and their final grid positions from one table; arrivals can land on a cue while the camera remains continuous.
+- Preserve real interface labels and supported states. Abstract panels are conceptual illustrations, not evidence of shipped features.
+
+The study is silent, original procedural art. Photographic studies use illustrated media cards and product studies use an illustrative object; replace these with actual supplied/generated assets for production. It demonstrates material, transforms and layout rather than proving real product claims. Configure an original subject, palette, scene sequence, duration and soundtrack from the brief.
+
+Use `src/studies/timing.ts` for deterministic easing, seeded variation, measured stroke truncation and cue helpers. `window.animaguru.texts` exposes visible study text bounds for `npm run review`; text declarations alone do not count as visual proof. Inspect the actual frames as well.
+
+## Sources and attribution
+
+Inspired by the following pinned references; source-specific 3D/media dependencies are not bundled. New study art and scene geometry are original. The workspace's `VIDEO-STYLE-NOTICES.txt` preserves the repository notices and clarifies third-party asset scope.
+
+- [styles/dark-keynote/STYLE.md](https://github.com/lemomo-ai/lemo-opuscar/blob/b65efe6ea393fdc71212898ea4c57c754cfefa46/styles/dark-keynote/STYLE.md)
+- [styles/dark-keynote/demo/engine.js](https://github.com/lemomo-ai/lemo-opuscar/blob/b65efe6ea393fdc71212898ea4c57c754cfefa46/styles/dark-keynote/demo/engine.js)
+- [prompts/howdevelop-733090.md](https://github.com/yihui-dev/awesome-opus5-5-videos/blob/3d54892e2ae5b0e8d337171e6508bba4cec01ab8/prompts/howdevelop-733090.md)

package/skills/animaguru-diagram-formula-motion/SKILL.md

@@ -0,0 +1,36 @@
+---
+name: animaguru-diagram-formula-motion
+description: Build animated diagrams, plotted geometry and equation transformations for visual explanation. Use for the Diagram and Formula Motion visual style in an AnimaGuru video workspace.
+---
+
+# Diagram & Formula Motion
+
+Make relationships visible through nodes, paths, geometry and symbols. Use the user's Explainer, Advertisement or General Video brief to determine the subject and depth; equations are optional.
+
+## Establish the direction
+
+Follow the sibling `animaguru-video` category → style → quick interview flow for a new film. Reuse supplied choices and existing briefs; honor an explicit request to skip the interview. Share [the style preview](https://animaguru.com/animation-styles#diagram-formula-motion). Include missing choices in the shared interview:
+
+- Which concept and source should guide the explanation, and what knowledge can viewers assume?
+- Prefer abstract diagrams, plotted geometry, equations, or a combination?
+- Which notation, color meanings and degree of visual detail should remain consistent?
+
+## Design and build
+
+- Read `pixijs`, graphics, math and text guidance. Identify the quantities, entities and relationships before animating them. Maintain a small visual vocabulary so the same color and shape retain their meaning.
+- Keep semantic objects stable across transitions: a node changes position but keeps its identity; an equation term becomes a labelled segment; a selected parameter changes the corresponding plot. Use explicit object IDs for these relationships.
+- Use native graphics for simple diagrams and graphs. For typeset mathematics, create local SVG assets with an appropriate math renderer and import them into the Pixi canvas. Verify that fonts or glyph outlines are included. DOM-only formula overlays will be missing from canvas captures.
+- Lay out equations as meaningful tokens or groups when individual terms must move. Preserve fraction rules, superscripts, signs and spacing. Ordinary text splitting does not understand mathematical structure.
+- Distinguish equality-preserving algebra, substitution, approximation and changes of representation. Use the correct signs and keep stated domains and units consistent. Validate formulas and graph samples numerically where possible; consult authoritative sources for unfamiliar claims.
+- Animate relationships at a readable pace: reveal a connection, move the relevant quantity, then allow the viewer to see the resolved state. Do not morph unrelated symbols in ways that imply a false equivalence.
+- Compute plots from their functions or supplied data. Choose domains that expose the intended behavior; split paths around discontinuities and undefined values. Use one world-to-screen mapping for geometry and labels.
+- AnimaGuru can provide authorized supporting illustrations and audio. Keep authoritative equations, values and diagrams authored and inspectable.
+- Drive all states and parameter changes from absolute timeline time and update resource intervals. Use the existing PixiJS player and export pipeline; additional rendering tools may prepare assets without replacing the player.
+
+## Review and completion
+
+Check each transformation's meaning as well as its appearance. Inspect formula baselines, label collisions, graph endpoints and discontinuities at transition boundaries and small viewing sizes. Match visual events to the real narration where present. Follow `animaguru-video` to deliver source expressions, diagram data, validation notes and the working HTML5 film; export MP4 when requested.
+
+## Additional visual treatments
+
+Read [the material and motion variants](references/variants.md) when the brief calls for a specialized treatment. Choose the relevant recipe without imposing its example story.

package/skills/animaguru-diagram-formula-motion/references/variants.md

@@ -0,0 +1,17 @@
+# Timed concept and simulation diagrams
+
+Choose one coherent treatment from the brief; these are visual/motion recipes, not mandatory stories. They apply to Explainer, Advertisement and General Video.
+
+## Semantic identity
+
+Give objects, terms and relationships stable IDs. A term entering a formula matches the object or variable it denotes; labels and connector endpoints follow geometry. Reveal enough context before changing a parameter.
+
+## Timed process simulation
+
+Define state from absolute time or bake a checked event schedule. Token buckets, flows or measured recipe quantities can be shown with this approach. Topic-specific physical/numerical truth needs independent source/calculation checks.
+
+## Reading and numerical checks
+
+Pause moving geometry when required text needs attention. Check domains, units, graph discontinuities and term transformations. Do not imply equality or causal relationships just because shapes animate smoothly into one another.
+
+Use the bundled absolute-time helpers in `src/studies/timing.ts` where appropriate. Keep actual resources and cues in the timeline, then inspect cold seeks, full-speed motion and exported frames. The shared video skill governs interviews, generation scope and delivery.

package/skills/animaguru-engraved-relief-print/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: animaguru-engraved-relief-print
+description: Build Engraved & Relief Print visuals and motion in an AnimaGuru PixiJS video workspace. Use when this visual style is chosen for an explainer, advertisement or general video.
+---
+
+# Engraved & Relief Print
+
+Form-following hatch marks and carved impressions revealed through stable mark schedules and layered print treatments. Subject, audience, emotional tone and story come from the user's brief. This style supports Explainer, Advertisement and General Video.
+
+## Establish direction
+
+Follow sibling `animaguru-video` for purpose → style → compact interview before new production. Reuse explicit choices and completed briefs; honor a requested interview skip. Share [the style gallery](https://animaguru.com/animation-styles#engraved-relief-print). Fold missing visual questions into the shared interview:
+
+- Which process fits: copperplate, dark relief or color woodblock?
+- How much fine detail and how many color plates are wanted?
+
+## Build the look and movement
+
+Read `pixijs` and the relevant Graphics, Text, Container, Assets or Mesh skills. Read `animaguru` before generating assets; discover current models and estimates, respect the user's budget, and reuse completed media. Essential text and exact numerical geometry belong in authored canvas content.
+
+- Choose a process before drawing. Copperplate uses swelling dark hatches; relief uses light gouges in dark masses; color woodblock uses separate keyed plates.
+- Hatching follows form and remains local to moving pieces. Cache the geometry, schedule reveal by stable mark IDs, and test line density at delivery scale.
+- Prepare underpainting/solid areas before fine marks so early frames read. Do not reduce all three processes to the same noise overlay.
+
+Use the [production recipe](references/recipe.md) for the bundled original study and port-specific decisions. It is a technique example, not a prescribed film. Keep motions a function of absolute seconds; separate drawing cadence from camera/light cadence when the material calls for it. Prepare assets locally, preserve async seeking and register actual resource/cue times. Keep essential type editable and localized.
+
+## Review and delivery
+
+Inspect form direction and mark survival after compression; stable hatch texture, paper margins and ink-layer order. Check representative full-resolution frames, moving playback, desktop/mobile fit and cold seeks. Follow `animaguru-video` for production records, audio mixing and HTML5 delivery; export MP4 when requested. Describe the actual 2D/2.5D rendering method accurately. Physically rendered 3D remains deferred.

package/skills/animaguru-engraved-relief-print/references/recipe.md

@@ -0,0 +1,24 @@
+# Engraved & Relief Print production recipe
+
+The workspace includes an executable original study in `src/studies/scenes.ts`, selected with `engraved-relief-print`. Its drawing commands are consumed by `src/studies/pixi-renderer.ts`; geometry and time sampling are shared with the gallery renderer. All copy lives in `shared/locales/en/style-studies.json`.
+
+In an untouched starter, run `npm run example -- engraved-relief-print`, then `npm run dev`. The command refuses to replace an authored film, retains starter backups in `production/example-backup/`, and writes an 8-second study with its own resource tracks. `npm run export:html` and `npm run export:mp4` use this installed film normally. To use the recipe inside an existing film, import `studyFrame()` and `StudyRenderer`, then compose the selected scene into the project's timeline instead of running the installer.
+
+## Adaptation decisions
+
+- Choose a process before drawing. Copperplate uses swelling dark hatches; relief uses light gouges in dark masses; color woodblock uses separate keyed plates.
+- Hatching follows form and remains local to moving pieces. Cache the geometry, schedule reveal by stable mark IDs, and test line density at delivery scale.
+- Prepare underpainting/solid areas before fine marks so early frames read. Do not reduce all three processes to the same noise overlay.
+
+The study is silent, original procedural art. Photographic studies use illustrated media cards and product studies use an illustrative object; replace these with actual supplied/generated assets for production. It demonstrates material, transforms and layout rather than proving real product claims. Configure an original subject, palette, scene sequence, duration and soundtrack from the brief.
+
+Use `src/studies/timing.ts` for deterministic easing, seeded variation, measured stroke truncation and cue helpers. `window.animaguru.texts` exposes visible study text bounds for `npm run review`; text declarations alone do not count as visual proof. Inspect the actual frames as well.
+
+## Sources and attribution
+
+Inspired by the following pinned references; source-specific 3D/media dependencies are not bundled. New study art and scene geometry are original. The workspace's `VIDEO-STYLE-NOTICES.txt` preserves the repository notices and clarifies third-party asset scope.
+
+- [styles/engraving/STYLE.md](https://github.com/lemomo-ai/lemo-opuscar/blob/b65efe6ea393fdc71212898ea4c57c754cfefa46/styles/engraving/STYLE.md)
+- [styles/engraving/demo/engine/burin.js](https://github.com/lemomo-ai/lemo-opuscar/blob/b65efe6ea393fdc71212898ea4c57c754cfefa46/styles/engraving/demo/engine/burin.js)
+- [styles/woodcut/STYLE.md](https://github.com/lemomo-ai/lemo-opuscar/blob/b65efe6ea393fdc71212898ea4c57c754cfefa46/styles/woodcut/STYLE.md)
+- [styles/ukiyoe/STYLE.md](https://github.com/lemomo-ai/lemo-opuscar/blob/b65efe6ea393fdc71212898ea4c57c754cfefa46/styles/ukiyoe/STYLE.md)

package/skills/animaguru-flat-vector-motion/SKILL.md

@@ -0,0 +1,35 @@
+---
+name: animaguru-flat-vector-motion
+description: Build animated videos from flat geometric illustrations, icons, connectors and shape transitions. Use for the Flat Vector Motion visual style in an AnimaGuru video workspace.
+---
+
+# Flat Vector Motion
+
+Create a coherent language of shapes and intentional movement. The brief supplies subject, story, audience and pacing; this style supports Explainer, Advertisement and General Video projects.
+
+## Establish the direction
+
+Use the sibling `animaguru-video` category → style → quick interview flow for new films. Reuse supplied answers and completed briefs; honor an explicit request to skip the interview. Share [the style preview](https://animaguru.com/animation-styles#flat-vector-motion). Fold missing choices into the shared interview:
+
+- Which palette, geometric language and illustration references match the brand?
+- Should movement feel precise and restrained, soft and friendly, or energetic?
+- Are icons or illustrations supplied, and are there shapes or visual metaphors to preserve or avoid?
+
+## Design and build
+
+- Read `pixijs` and its graphics, container and text guidance. Set shared stroke widths, corner radii, proportions and spacing before creating scenes. Keep perspective, icon detail and illustration weight consistent.
+- Build simple geometry in PixiJS or SVG. Use AnimaGuru's `vector` style for complex illustrations when useful; inspect the current catalog and generation guidance. Generated vector-looking images remain bitmaps unless the delivered file is actually vector. Keep critical labels editable.
+- Compose with a clear focal element and supporting shapes. Use size and whitespace to distinguish information levels. Choose flat fills and limited shading according to the brief.
+- Connect scenes through shared shapes or layout anchors: a circle becomes a node, a card expands into a panel, or a connector guides the next reveal. Avoid arbitrary transitions unrelated to the composition.
+- Use anticipation, travel and settling deliberately. Stagger related elements while allowing each composition to resolve. Motion should explain grouping, direction or emphasis.
+- Interpolate authored geometry only when topology and point ordering match. For unrelated silhouettes, use a mask, matched transform or crossfade rather than an unstable point-to-point morph. Check the intermediate silhouettes.
+- Cache static drawing where useful and animate transforms. Rebuild geometry only when its shape actually changes; do not regenerate assets during playback. Keep effects within the portable player.
+- Follow the shared deterministic timeline and resource metadata contract. Derive connector progress, masks and transforms from absolute time, including on backward seeks.
+
+## Review and completion
+
+Follow `animaguru-video` for source records, audio and export. Check outline consistency, seam-free masks, transition midpoints and readable holds at the intended viewing size. Confirm that essential meaning survives a still frame and silent playback when required by the brief. Deliver consistent vector styling, editable text, synchronized resources and a working HTML5 film; export MP4 when requested.
+
+## Additional visual treatments
+
+Read [the material and motion variants](references/variants.md) when the brief calls for a specialized treatment. Choose the relevant recipe without imposing its example story.

package/skills/animaguru-flat-vector-motion/references/variants.md

@@ -0,0 +1,17 @@
+# Vector performance variants
+
+Choose one coherent treatment from the brief; these are visual/motion recipes, not mandatory stories. They apply to Explainer, Advertisement and General Video.
+
+## Pictogram acting
+
+Store joint poses and props in data; choose one clearly readable action per visual focus. Use a consistent figure scale, stroke family and geometry grid. A cue table can schedule contacts and releases without prescribing a numbered sports or instruction story.
+
+## Hard-shadow panels
+
+Use opaque flat shapes and a deliberate offset silhouette for depth. Keep shadow direction and width consistent. Layer transitions can use grid bands, matched shapes or occlusion rather than swapping whole unrelated slides.
+
+## Rhythm and microgame panels
+
+Call/response, round cards, timers and replay are optional workflow tools. If used, drive them from the same cue table and make each action/result readable. Do not force escalating difficulty, lives or a fixed frame-rate ladder on every vector film.
+
+Use the bundled absolute-time helpers in `src/studies/timing.ts` where appropriate. Keep actual resources and cues in the timeline, then inspect cold seeks, full-speed motion and exported frames. The shared video skill governs interviews, generation scope and delivery.

package/skills/animaguru-general-video/SKILL.md

@@ -0,0 +1,20 @@
+---
+name: animaguru-general-video
+description: Develop and produce original films, music visuals, titles, ambient loops and visual essays in an AnimaGuru video workspace. Use after the General Video purpose is selected.
+---
+
+# AnimaGuru general video
+
+Read sibling `animaguru-video` and the selected visual skill. Purpose describes the video's job; visual style remains independently selectable. For new work, run the shared compact interview once and fold workflow-specific missing questions into it. Reuse explicit answers, saved project choices and completed briefs; honor a requested interview skip.
+
+Infer a suitable workflow from the request or clarify it in that interview. Do not add a mandatory menu when the user's intent is clear. Save a listed workflow ID in `animaguru.project.json` only when it fits; leave it null for a custom brief. Read only the applicable reference:
+
+- **Original Short Film** (`short-film`): Create a film from a subject, character or idea with an intentional ending. Read [the workflow](references/short-film.md).
+- **Music & Lyric Video** (`music-lyrics`): Choreograph visual changes to the actual track, lyrics and sections. Read [the workflow](references/music-lyrics.md).
+- **Titles & Event Motion** (`titles-event`): Present names, credits or event identity in a composed sequence. Read [the workflow](references/titles-event.md).
+- **Ambient & Seamless Loop** (`ambient-loop`): Build a mood or display loop with continuous motion at its boundary. Read [the workflow](references/ambient-loop.md).
+- **Visual Essay & Podcast** (`visual-essay`): Add accurate graphics and chapters to source text or recorded speech. Read [the workflow](references/visual-essay.md).
+
+Develop an original sequence from the user's sources, audience, tone, duration and desired outcome. Avoid importing fixed scene counts, mandatory emotional arcs, default audiences or a compulsory CTA. Respect supplied narration/music preferences. Discover AnimaGuru generation models and costs through the local main skill; this purpose selection does not authorize paid jobs.
+
+Use `src/studies/media.ts` for optional deterministic source-time/trim mapping and cue comparison; prepare footage locally and await decoded frames before capture. `npm run review` collects actual canvas text bounds, deterministic seek comparisons and review frames through the browser. Preserve resource timing and the HTML5 embed/export contract. Review moving playback and audio as well as numerical checks.

package/skills/animaguru-general-video/references/ambient-loop.md

@@ -0,0 +1,11 @@
+# Ambient & Seamless Loop
+
+Build a mood or display loop with continuous motion at its boundary.
+
+Fold this missing information into the shared quick interview: **What mood, loop length, audio, motion tolerance and display context should the loop serve?** Reuse supplied answers. Save the chosen angle, constraints and evidence in `production/brief.md`; a workflow is not a fixed story template.
+
+- Choose a periodic time model or a fully occluded handoff. Check position, opacity, light, particles, source-video time and movement velocity at the boundary.
+- Audio loops need their own seam check; a silent visual loop does not establish soundtrack continuity. Derive procedural variation from periodic functions or a baked loop schedule.
+- Keep display and embed behavior appropriate: autoplay muted where supported, preserve an accessible way to start/unmute, and honor reduced-motion preferences.
+
+Completion: First/last image and velocity are continuous, audio seams are clean, and embedded playback has working controls. Use `animaguru-video` for the common production, budget, review and HTML5/MP4 contract.

package/skills/animaguru-general-video/references/music-lyrics.md

@@ -0,0 +1,11 @@
+# Music & Lyric Video
+
+Choreograph visual changes to the actual track, lyrics and sections.
+
+Fold this missing information into the shared quick interview: **Which actual track and lyrics, section boundaries and imagery should drive the film?** Reuse supplied answers. Save the chosen angle, constraints and evidence in `production/brief.md`; a workflow is not a fixed story template.
+
+- Analyze the real track rather than assuming a tempo. Save measured downbeats/section/energy cues or derive them from an original cue-based composition.
+- Choose whether words follow vocal phrasing, whether cuts follow music, and where visuals should hold. Phrase timing and instrumental beats need not coincide.
+- Bake audio envelopes and particle schedules. Match the visible lyrics to actual recorded words; do not rely on live AudioContext state for export.
+
+Completion: Lyrics are accurate, cues align, and both silent visual reading and full-speed audio playback work. Use `animaguru-video` for the common production, budget, review and HTML5/MP4 contract.

package/skills/animaguru-general-video/references/short-film.md

@@ -0,0 +1,11 @@
+# Original Short Film
+
+Create a film from a subject, character or idea with an intentional ending.
+
+Fold this missing information into the shared quick interview: **What subject, tone, characters, sound/dialogue and ending intent should guide the film?** Reuse supplied answers. Save the chosen angle, constraints and evidence in `production/brief.md`; a workflow is not a fixed story template.
+
+- Write a treatment and shot choices appropriate to the subject. A hook, a turn, dialogue or a four-camera-move quota is not universally required.
+- Give meaningful actions anticipation, action and follow-through. Maintain prop contact, eyelines and spatial relationships through camera moves.
+- Use the selected style as material grammar. Sound may be narration, dialogue, ambience, foley, music or silence according to the brief.
+
+Completion: Readable performance, intentional pacing, coherent space and an ending that serves the brief. Use `animaguru-video` for the common production, budget, review and HTML5/MP4 contract.

package/skills/animaguru-general-video/references/titles-event.md

@@ -0,0 +1,11 @@
+# Titles & Event Motion
+
+Present names, credits or event identity in a composed sequence.
+
+Fold this missing information into the shared quick interview: **What exact names, credit hierarchy, event facts and output formats must be presented?** Reuse supplied answers. Save the chosen angle, constraints and evidence in `production/brief.md`; a workflow is not a fixed story template.
+
+- Keep names and credit roles in localized structured data. Confirm spelling and hierarchy before designing letterforms or motion.
+- Choose a camera and transition grammar from the style. Readable timing overrides a desire to animate every name on a beat.
+- Subjects can be events, films, shows or original identities. Do not inherit the source example's spy/thriller plot, fictional institution or credit sequence.
+
+Completion: Names, dates and roles are correct; every important title fully arrives and holds long enough. Use `animaguru-video` for the common production, budget, review and HTML5/MP4 contract.

package/skills/animaguru-general-video/references/visual-essay.md

@@ -0,0 +1,11 @@
+# Visual Essay & Podcast
+
+Add accurate graphics and chapters to source text or recorded speech.
+
+Fold this missing information into the shared quick interview: **Which source text/audio, speakers and chapter priorities should the film preserve?** Reuse supplied answers. Save the chosen angle, constraints and evidence in `production/brief.md`; a workflow is not a fixed story template.
+
+- Map source sentences and ideas to visual cues. Add graphics when they explain a relationship, illustrate evidence or clarify the speaker's point.
+- Keep graphics inside coherent source-audio stretches. Preserve factual meaning and quotation accuracy; PiP is optional and follows the requested speaker visibility.
+- Use a shared style bible and component kit for repeated inserts. Avoid importing a fixed animation count, compulsory camera zooms or a blanket multi-agent judging process.
+
+Completion: Faithful source meaning, cue synchronization, consistent insert design and clear chapter flow. Use `animaguru-video` for the common production, budget, review and HTML5/MP4 contract.

package/skills/animaguru-generative-light-signal/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: animaguru-generative-light-signal
+description: Build Generative Light & Signal visuals and motion in an AnimaGuru PixiJS video workspace. Use when this visual style is chosen for an explainer, advertisement or general video.
+---
+
+# Generative Light & Signal
+
+Luminous traces, geometric fields and seeded paths choreographed by time or precomputed audio cues. Subject, audience, emotional tone and story come from the user's brief. This style supports Explainer, Advertisement and General Video.
+
+## Establish direction
+
+Follow sibling `animaguru-video` for purpose → style → compact interview before new production. Reuse explicit choices and completed briefs; honor a requested interview skip. Share [the style gallery](https://animaguru.com/animation-styles#generative-light-signal). Fold missing visual questions into the shared interview:
+
+- Should geometry feel signal-like, organic or kaleidoscopic?
+- Is audio reactivity needed, and what motion intensity is comfortable?
+
+## Build the look and movement
+
+Read `pixijs` and the relevant Graphics, Text, Container, Assets or Mesh skills. Read `animaguru` before generating assets; discover current models and estimates, respect the user's budget, and reuse completed media. Essential text and exact numerical geometry belong in authored canvas content.
+
+- Choose one coherent recipe and precompute audio beat/energy envelopes if needed. Export must not depend on a live analyser or browser sound playback.
+- Compute particle paths analytically or from a baked schedule and keep geometry within an explicit budget. Stable identities improve both seeking and comprehension.
+- Effect intensity follows the brief; avoid compulsory flashing, rainbow hues and permanent distortion. Preserve clear type regions when copy is required.
+
+Use the [production recipe](references/recipe.md) for the bundled original study and port-specific decisions. It is a technique example, not a prescribed film. Keep motions a function of absolute seconds; separate drawing cadence from camera/light cadence when the material calls for it. Prepare assets locally, preserve async seeking and register actual resource/cue times. Keep essential type editable and localized.
+
+## Review and delivery
+
+Inspect measured cue alignment and random-order frame stability; effect intensity, contrast and low-powered-device playback. Check representative full-resolution frames, moving playback, desktop/mobile fit and cold seeks. Follow `animaguru-video` for production records, audio mixing and HTML5 delivery; export MP4 when requested. Describe the actual 2D/2.5D rendering method accurately. Physically rendered 3D remains deferred.

package/skills/animaguru-generative-light-signal/references/recipe.md

@@ -0,0 +1,23 @@
+# Generative Light & Signal production recipe
+
+The workspace includes an executable original study in `src/studies/scenes.ts`, selected with `generative-light-signal`. Its drawing commands are consumed by `src/studies/pixi-renderer.ts`; geometry and time sampling are shared with the gallery renderer. All copy lives in `shared/locales/en/style-studies.json`.
+
+In an untouched starter, run `npm run example -- generative-light-signal`, then `npm run dev`. The command refuses to replace an authored film, retains starter backups in `production/example-backup/`, and writes an 8-second study with its own resource tracks. `npm run export:html` and `npm run export:mp4` use this installed film normally. To use the recipe inside an existing film, import `studyFrame()` and `StudyRenderer`, then compose the selected scene into the project's timeline instead of running the installer.
+
+## Adaptation decisions
+
+- Choose one coherent recipe and precompute audio beat/energy envelopes if needed. Export must not depend on a live analyser or browser sound playback.
+- Compute particle paths analytically or from a baked schedule and keep geometry within an explicit budget. Stable identities improve both seeking and comprehension.
+- Effect intensity follows the brief; avoid compulsory flashing, rainbow hues and permanent distortion. Preserve clear type regions when copy is required.
+
+The study is silent, original procedural art. Photographic studies use illustrated media cards and product studies use an illustrative object; replace these with actual supplied/generated assets for production. It demonstrates material, transforms and layout rather than proving real product claims. Configure an original subject, palette, scene sequence, duration and soundtrack from the brief.
+
+Use `src/studies/timing.ts` for deterministic easing, seeded variation, measured stroke truncation and cue helpers. `window.animaguru.texts` exposes visible study text bounds for `npm run review`; text declarations alone do not count as visual proof. Inspect the actual frames as well.
+
+## Sources and attribution
+
+Inspired by the following pinned references; source-specific 3D/media dependencies are not bundled. New study art and scene geometry are original. The workspace's `VIDEO-STYLE-NOTICES.txt` preserves the repository notices and clarifies third-party asset scope.
+
+- [prompts/gandamu-ml-220530.md](https://github.com/yihui-dev/awesome-opus5-5-videos/blob/3d54892e2ae5b0e8d337171e6508bba4cec01ab8/prompts/gandamu-ml-220530.md)
+- [prompts/pound75423-464968.md](https://github.com/yihui-dev/awesome-opus5-5-videos/blob/3d54892e2ae5b0e8d337171e6508bba4cec01ab8/prompts/pound75423-464968.md)
+- [prompts/rneayan-406275.md](https://github.com/yihui-dev/awesome-opus5-5-videos/blob/3d54892e2ae5b0e8d337171e6508bba4cec01ab8/prompts/rneayan-406275.md)

package/skills/animaguru-glass-mosaic/SKILL.md

@@ -0,0 +1,29 @@
+---
+name: animaguru-glass-mosaic
+description: Build Glass Mosaic visuals and motion in an AnimaGuru PixiJS video workspace. Use when this visual style is chosen for an explainer, advertisement or general video.
+---
+
+# Glass Mosaic
+
+Colored panes or loose glass-like tiles, strong seams and moving transmitted-light treatments in 2.5D. Subject, audience, emotional tone and story come from the user's brief. This style supports Explainer, Advertisement and General Video.
+
+## Establish direction
+
+Follow sibling `animaguru-video` for purpose → style → compact interview before new production. Reuse explicit choices and completed briefs; honor a requested interview skip. Share [the style gallery](https://animaguru.com/animation-styles#glass-mosaic). Fold missing visual questions into the shared interview:
+
+- Should panes stay joined or should loose tiles move?
+- Which border material, color range and light movement fit?
+
+## Build the look and movement
+
+Read `pixijs` and the relevant Graphics, Text, Container, Assets or Mesh skills. Read `animaguru` before generating assets; discover current models and estimates, respect the user's budget, and reuse completed media. Essential text and exact numerical geometry belong in authored canvas content.
+
+- Store tiles as stable semantic pieces with shared edges; fixed-pane and loose-tile variants have different motion rules.
+- Use authored lighting masks and projected tilts in the 2.5D treatment. Physical refraction and arbitrary thick-tile orbit remain deferred.
+- Preserve hue and seam readability as light changes. Tile motion must expose prepared gaps/backing rather than inventing hidden imagery.
+
+Use the [production recipe](references/recipe.md) for the bundled original study and port-specific decisions. It is a technique example, not a prescribed film. Keep motions a function of absolute seconds; separate drawing cadence from camera/light cadence when the material calls for it. Prepare assets locally, preserve async seeking and register actual resource/cue times. Keep essential type editable and localized.
+
+## Review and delivery
+
+Inspect stable tile ids and coherent shared seams; color saturation and readable silhouettes through light changes. Check representative full-resolution frames, moving playback, desktop/mobile fit and cold seeks. Follow `animaguru-video` for production records, audio mixing and HTML5 delivery; export MP4 when requested. Describe the actual 2D/2.5D rendering method accurately. Physically rendered 3D remains deferred.