[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6BXZNFJZXG84J91AWC5K5JY":3,"skill-same:01M6BXZNFJZXG84J91AWC5K5JY":73},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":39,"linkOnly":40,"markdownContent":41,"markdownBody":42,"frontMatter":43,"refName":44,"commitSha":45,"skillDirectory":46,"forksCount":47,"repositoryHtmlUrl":48,"repositoryHomepageUrl":49,"ownerHtmlUrl":50,"ownerBlogUrl":12,"ownerDescription":51,"updatedAt":52,"install":53,"resources":66},"01M6BXZNFJZXG84J91AWC5K5JY","frame-composition",null,"Stage the frame on a NodeTool timeline — grids, focal placement, safe areas per aspect ratio, depth layers and parallax, camera moves, and where elements enter and leave. Use when placing titles, lower thirds, shapes or overlays, when a layout must survive 16:9 and 9:16, when adding a push or a parallax move, or when a frame reads flat or cramped. Not for what a clip says or when it cuts.","nodetool-ai\u002Fnodetool","nodetool-ai","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F164873569?v=4","01M6BXPFDW36K9K1MR8ZH63HG2","nodetool.ai","01M6BXPFJ893XM219GQN60B0WK",635,"AGPL-3.0","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fframe-composition\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38],"Composition","frame composition","frame","layout","when","depth","parallax","for","overlays","layers","move","elements","safe","stage","camera","clip","use","nodetool","timeline","push","2026-11-10 10:12:30",false,"---\nname: frame-composition\ndescription: Stage the frame on a NodeTool timeline — grids, focal placement, safe areas per aspect ratio, depth layers and parallax, camera moves, and where elements enter and leave. Use when placing titles, lower thirds, shapes or overlays, when a layout must survive 16:9 and 9:16, when adding a push or a parallax move, or when a frame reads flat or cramped. Not for what a clip says or when it cuts.\n---\n\n# Frame Composition → staging on the timeline\n\nWhere things sit, how deep the frame is, and how the camera moves through it.\n`motion-graphics` is the build loop; `timeline-edit-ops` carries the op\ncontract. This decides the coordinates you pass either.\n\n## Coordinate spaces, and mixing them is the usual bug\n\n| What | Space | Read as |\n|---|---|---|\n| `shapeStyle.x\u002Fy\u002Fwidth\u002Fheight`, `x2\u002Fy2`, `d` | 0..1 of the **frame** | 0.5 is the frame's middle |\n| `set_mask` `x\u002Fy\u002Fwidth\u002Fheight`, `d` | 0..1 of the **layer** | 0.5 is the clip's own middle, not the frame's |\n| `positionX` \u002F `positionY` curves | canvas px | Absolute, and folds as `replace` |\n| `offsetX` \u002F `offsetY` curves | canvas px, added | Composes with everything else |\n| `slide` `distance`, `kenBurns` `driftX\u002FdriftY` | fraction of frame width or height | 0.3 travels 30% of the frame |\n| `anchorX` \u002F `anchorY` | 0..1 | The point a scale or a rotation pivots on |\n| `textStyle.fontSizePx` | sequence px | Against `height` from `get_timeline`, not the preview width |\n| `transform.depthPx`, `camera2d.position\u002FdepthPx\u002FfocalLengthPx` | sequence px | Perspective; clip depth also orders clips within one track |\n| `layout.gap`, `layout.padding` | sequence px | Flex container spacing, measured against its children's rendered boxes |\n\nRead `width`, `height` and `fps` off `get_timeline` before you compute anything.\nA layout authored against 1920×1080 and saved onto a 1080×1920 sequence is off\nby more than a crop.\n\n## Text is placed by containers\n\nBuilding with `@nodetool-ai\u002Fsandbox-timeline` (see `motion-graphics` for when\ncode is the right surface)? **A container decides where its children sit —\nyou almost never write `x`\u002F`y` for text.** `s.stack(children, o)` and\n`s.row(children, o)` build a real Yoga flex container and reparent\n`children` onto it as its real children — nest freely, a `row` of `stack`s\nor a `stack` holding a nested `row` both resolve correctly. A text child\nreflows to whatever width Yoga offers it, not a guessed line length.\n\nThe container's own placement in the frame is `o.at` (`{x, y}`, default the\nframe centre) plus `o.anchor` (default `\"center\"`): `anchor` names the point\nof the container's own *computed* box that lands at `at` — the name\n`\"left\"` puts that box's left edge at `at.x`, no measuring a sibling's width\nrequired. Named anchors: `\"top-left\"`, `\"top\"`, `\"top-right\"`, `\"left\"`,\n`\"center\"`, `\"right\"`, `\"bottom-left\"`, `\"bottom\"`, `\"bottom-right\"`, or pass\n`{x, y}` fractions directly. `o.align` (cross axis, default `\"start\"`) and\n`o.justify` (main axis, unset) are the Yoga `alignItems`\u002F`justifyContent`.\nA child's own flex sizing is `o.flexItem` on that child's own call (`grow`,\n`width`, `alignSelf`, …); `absolute: true` on a `rect`\u002F`ellipse` (or any\nclip) takes it out of flow and resizes it to fill its flex parent's box\nexactly — how a plate sits behind a card's text with no pixel math.\n\nReach for explicit `x`\u002F`y` only for decoration (a stray mark, a background\nshape) or a deliberate free placement that has nothing to do with reading\norder. Anything a viewer reads — a title, a kicker, a caption, a row of\nstats — goes in a container.\n\n**Title block** — kicker, title, sub, stacked with a shared rhythm:\n\n```js\n\u002F\u002F fragment\nconst kicker = s.kicker(\"LAUNCHING TODAY\");\nconst title = s.text(\"Round-ups\", { size: 120, weight: 700 });\nconst sub = s.text(\"on every card\", { size: 40 });\ns.stack([kicker, title, sub], { gap: 20 });\n```\n\n**Lockup** — a mark beside its wordmark, one row:\n\n```js\n\u002F\u002F fragment\nconst mark = s.image(logoAssetId);\nconst wordmark = s.text(\"Northwind\", { size: 64, weight: 700 });\ns.row([mark, wordmark], { gap: 16, align: \"center\" });\n```\n\n**Card** — a padded column over an absolute plate:\n\n```js\n\u002F\u002F fragment\nconst plate = s.rect(1, 1, \"#0c2a21\", { r: 24, absolute: true });\nconst heading = s.text(\"Plan Pro\", { size: 48, weight: 700 });\nconst price = s.text(\"$12\u002Fmo\", { size: 32 });\ns.stack([plate, heading, price], { gap: 8, padding: 24 });\n```\n\n**List \u002F grid** — a row that wraps into a grid when it runs out of width:\n\n```js\n\u002F\u002F fragment\nconst chips = [\"Round-ups\", \"Instant transfers\", \"No fees\"].map((t) => s.pill(t));\ns.row(chips, { gap: 12, wrap: true });\n```\n\n## Grid\n\nTwelve columns for landscape, six for vertical. For a 1920-wide frame with a\n100px outer margin and a 24px gutter:\n\n```\ncontent = 1920 − 2×100 = 1720\ncolumn  = (1720 − 11×24) \u002F 12 = 121.3px\n```\n\nSpace everything in multiples of 8. `fontSizePx` for body copy sits near 2.5–4%\nof frame height; below 2.5% `validate_timeline` reports `text_illegible`.\nChoose the family, weights, and title scale from `motion-direction` § Typography\nbefore placing text on this grid.\n\n## One focal point\n\nPlace the subject on a thirds power point, not dead centre, unless the piece\nuses symmetry deliberately. The intersections:\n\n| Sequence | Power points (px) |\n|---|---|\n| 1920×1080 | x 640 \u002F 1280, y 360 \u002F 720 |\n| 1080×1920 | x 360 \u002F 720, y 640 \u002F 1280 |\n| 1080×1080 | x 360 \u002F 720, y 360 \u002F 720 |\n\nRank the eye's path by **size, then contrast, then colour, then position**. One\nprimary per frame; everything else supports. Leave negative space around it —\na frame filled edge to edge reads as cheap and gives the motion nowhere to go.\n\n## Depth: planes and track order\n\nSplit the picture into background, midground and foreground. Track index\ncontrols order **between** tracks: lowest index draws on top. Within one\ntrack, clips sort by `transform.depthPx`: farther values draw first and\nnearer, positive values draw later, in front. The same depth controls\nperspective under the sequence's `camera2d`. Moving a clip to another track\nchanges its overlap order but not its camera distance.\n\n| Plane | Suggested track index | Clip depth | Motion and look |\n|---|---|---|---|\n| Foreground | lowest | Nearer, positive | The move the eye follows; sharp unless focus shifts |\n| Midground | middle | Around zero | The subject and contrast peak |\n| Background | highest | Farther, negative | Small drift and lower contrast |\n\nFor flat parallax, run three speeds over one window. One `kenBurns` per layer\nwith `zoom: 0` and a different drift is a `fullClip` preset, so all three\ncover the same span. Put the three on clips of the same length: the drift is\nspread over the clip, so its speed scales with clip length:\n\n```json\n{\"role\": \"loop\", \"preset\": \"kenBurns\",\n \"params\": {\"zoom\": 0, \"driftX\": -0.02, \"driftY\": 0}}\n```\n\nBackground −0.02, midground −0.06, foreground −0.12. One layer at one speed is\nnot parallax, it is the template moving.\n\nFor a shared 2.5D move, give the layers distinct `transform.depthPx` values and\nkeyframe the sequence's `camera2d`. Its absolute `timeMs` keyframes interpolate\ncamera position and depth linearly. `focusDepthPx` and `aperturePx` control depth of\nfield. Use one camera move to connect the planes, then inspect a middle frame\nfor unwanted scale or blur. Set track indexes for overlap between tracks and\nclip depth for order within a track.\n\nFor a tilted card wall, parent the cards to a group with the wall's 2D\n`rotation`, `rotationX`, and `perspective`. Give each card its own `depthPx`\nand scroll curve, then use `camera2d` for the shared push and depth blur.\nThis is a flat card projection: preview the wall at several frames to check\nthat spacing, overlap, and blur read as one assembly.\n\nFor a move with a shape of its own, write a `custom` curve instead and mind the\nrole: an `in` window holds its `t=0` values before it and contributes nothing\nafter, an `out` window holds its `t=1` values after it, and a `loop` repeats its\ncycle — so a one-way travel that must stay where it landed is an `out`, and a\n`loop` curve has to return to the value it started on or it jumps every cycle.\n\nParent the assembly to a group when the layers must hold their relationship\nthrough a move: `add_group`, then animate the group. Children keep their own\ntracks, so grouping never changes what covers what.\nFor a transition on the assembled picture, set `transitionIn` on the group. The\nchildren stay parented on their own tracks while the transition acts once on\ntheir precomposed surface.\n\nShape geometry is rasterized inside a frame-sized source surface before the\nclip is placed. Keep normalized `shapeStyle` geometry inside that surface,\nthen move the clip or parent group with `transform.position` or custom\n`offsetX`\u002F`offsetY` curves for scrolling walls and flying cards. Geometry\nauthored outside the source surface clips before placement.\n\n## Camera moves\n\nChoose between a sequence camera move and a local clip move. `camera2d`\nkeyframes move the view across all depth planes. A layer animation moves one\nclip, and its `anchor` decides where scale or rotation pivots.\n\n| Move | How | Timing |\n|---|---|---|\n| Push \u002F pull | `kenBurns` with `zoom` 0.08–0.2 and `direction` in or out | 1500–4000ms, `easeInOut` |\n| Drift | `kenBurns` `driftX` \u002F `driftY`, ±0.02–0.08 | Same window as the zoom |\n| Pan | `custom` curve on `offsetX` across the clip | 800–2000ms, `easeInOut` |\n| Whip | `offsetX` over 150–250ms with a `blur` effect at 12–24px | On a cut, never mid-shot |\n| Follow | `offsetX` and `offsetY` tracking the subject, subject held off-centre and leading | Whatever the action takes |\n\n`camera2d` is a document field. `set_clip_params` changes a clip's `transform`\nbut refuses `camera2d`, `layout`, `repeater`, `motionBlur`, and other new document\nfields. Read the full document with `get_timeline`, change the relevant fields,\nthen write the full document with `set_timeline_document`. Keep its tracks,\nclips, markers, tempo, setup, media tracks, template identity, and other fields\nin the write. `timeline-edit-ops` carries the\nop contract.\n\n`kenBurns` is `fullClip`: it ignores `durationMs` and `delayMs` and runs across\nthe whole clip. To time a push to something shorter, write it as a `custom`\n`scale` curve instead.\n\nOne move per beat. A push plus a pan plus a rotate reads as chaos, and the three\nfold together into a direction nobody chose.\n\n## Enter and exit\n\nMotion that follows reading direction feels natural; reversing it reads as going\nback. Match `slide`'s `direction` to intent — it names the edge the element\narrives **from**:\n\n- New content: from the right, or from below.\n- Dismissing or undoing: reverse the entrance.\n- Drilling in: `pop`, or a scale curve rather than a travel.\n\nEnter and leave through the nearest edge. Keep `distance` at or below 0.35 or\nthe travel reads as a slide of the whole template — pair a longer move with a\nscale or opacity change on the same window. Same vocabulary on the cut:\n`set_transition` `direction` on `wipe`, `push` and `slide` also names the edge\nthe incoming clip arrives from.\n\n## Safe areas\n\nKeep titles, logos, faces and calls to action inside the margin. The numbers are\nfractions of the frame, so convert against `width` and `height`.\n\n| Aspect | Top | Bottom | Sides |\n|---|---|---|---|\n| 16:9 | 5% | 5% | 5% |\n| 9:16 | 12–14% | 18–20% | 6% |\n| 1:1 | 7% | 7% | 7% |\n| 4:5 | 8% | 8% | 6% |\n\nOn 1080×1920 that leaves the usable band roughly y 250 to y 1570. A caption clip\nalready sits at `bottomMarginFrac` 0.12 by default, which is inside the phone UI\nzone on some platforms — raise it rather than assuming.\n\n## One layout, several aspects\n\nMaster in the widest target with the focal point inside the centre band, then\n**restack** for vertical: move the headline above the subject, pull the subject\nup out of the bottom UI zone, and raise `fontSizePx` by about 20%. Cropping a\n16:9 layout to 9:16 clips whatever sat in the outer columns, which is usually the\nlogo and the call to action.\n\nBuilding in code, restacking is usually one option change on the container\nabove — flip `s.row` to `s.stack` (or its `align`\u002F`justify`), or move it to a\nnew `at`\u002F`anchor` — not a rebuild of every child's `x`\u002F`y`.\n\nOn the ops path, `layout: {display: \"flex\", flexDirection: \"row\" | \"column\",\ngap}` on a group clip keeps its real children (their `parentId` names the\ngroup) laid out with ordinary CSS flexbox — `justifyContent`\u002F`alignItems`\nposition them, `padding` insets the container. For a shape behind text, give\nthe shape `flexItem: {position: \"absolute\", inset: 0}` inside a padded flex\ncolumn holding the text: the shape's box resizes to the text's padded box,\nincluding an animated font size, with no manual pixel math. These relations\navoid redoing pixel math for each text change. A 16:9 document still needs a\nseparate placement pass for 9:16, with safe areas checked at that size.\n\n## Check the frame, not the document\n\n`preview_timeline_frame` composites the real picture — every track in order,\ntransforms and opacity applied, animations mid-flight. Read the layer list it\nreturns beside each frame: it names the stack top first with each layer's\n`z_index`, `opacity` and wipe progress, which answers most staging questions\nwithout looking at a pixel.\n\nPreview at the midpoint of every move, not at its ends. What frames catch and a\nvalidator cannot: a title that slid outside the frame, a scrim covering the face\nit was meant to sit beside, a lower third under the wrong element because two\ntrack indexes are the wrong way round, a layer that never draws because it is on\nan audio track.\n\n## Before you finish\n\n- One focal point, on a power point or deliberately centred.\n- Every element assigned a track for overlap order and a depth plane when the camera moves.\n- One camera move or eased layer move, 800–4000ms.\n- Entrances through the nearest edge, `distance` ≤ 0.35 or paired with a second\n  channel.\n- Critical content inside the safe margin for every target aspect, vertical\n  restacked rather than cropped.\n- Previewed at the midpoints, and the layer stack read.\n","# Frame Composition → staging on the timeline\n\nWhere things sit, how deep the frame is, and how the camera moves through it.\n`motion-graphics` is the build loop; `timeline-edit-ops` carries the op\ncontract. This decides the coordinates you pass either.\n\n## Coordinate spaces, and mixing them is the usual bug\n\n| What | Space | Read as |\n|---|---|---|\n| `shapeStyle.x\u002Fy\u002Fwidth\u002Fheight`, `x2\u002Fy2`, `d` | 0..1 of the **frame** | 0.5 is the frame's middle |\n| `set_mask` `x\u002Fy\u002Fwidth\u002Fheight`, `d` | 0..1 of the **layer** | 0.5 is the clip's own middle, not the frame's |\n| `positionX` \u002F `positionY` curves | canvas px | Absolute, and folds as `replace` |\n| `offsetX` \u002F `offsetY` curves | canvas px, added | Composes with everything else |\n| `slide` `distance`, `kenBurns` `driftX\u002FdriftY` | fraction of frame width or height | 0.3 travels 30% of the frame |\n| `anchorX` \u002F `anchorY` | 0..1 | The point a scale or a rotation pivots on |\n| `textStyle.fontSizePx` | sequence px | Against `height` from `get_timeline`, not the preview width |\n| `transform.depthPx`, `camera2d.position\u002FdepthPx\u002FfocalLengthPx` | sequence px | Perspective; clip depth also orders clips within one track |\n| `layout.gap`, `layout.padding` | sequence px | Flex container spacing, measured against its children's rendered boxes |\n\nRead `width`, `height` and `fps` off `get_timeline` before you compute anything.\nA layout authored against 1920×1080 and saved onto a 1080×1920 sequence is off\nby more than a crop.\n\n## Text is placed by containers\n\nBuilding with `@nodetool-ai\u002Fsandbox-timeline` (see `motion-graphics` for when\ncode is the right surface)? **A container decides where its children sit —\nyou almost never write `x`\u002F`y` for text.** `s.stack(children, o)` and\n`s.row(children, o)` build a real Yoga flex container and reparent\n`children` onto it as its real children — nest freely, a `row` of `stack`s\nor a `stack` holding a nested `row` both resolve correctly. A text child\nreflows to whatever width Yoga offers it, not a guessed line length.\n\nThe container's own placement in the frame is `o.at` (`{x, y}`, default the\nframe centre) plus `o.anchor` (default `\"center\"`): `anchor` names the point\nof the container's own *computed* box that lands at `at` — the name\n`\"left\"` puts that box's left edge at `at.x`, no measuring a sibling's width\nrequired. Named anchors: `\"top-left\"`, `\"top\"`, `\"top-right\"`, `\"left\"`,\n`\"center\"`, `\"right\"`, `\"bottom-left\"`, `\"bottom\"`, `\"bottom-right\"`, or pass\n`{x, y}` fractions directly. `o.align` (cross axis, default `\"start\"`) and\n`o.justify` (main axis, unset) are the Yoga `alignItems`\u002F`justifyContent`.\nA child's own flex sizing is `o.flexItem` on that child's own call (`grow`,\n`width`, `alignSelf`, …); `absolute: true` on a `rect`\u002F`ellipse` (or any\nclip) takes it out of flow and resizes it to fill its flex parent's box\nexactly — how a plate sits behind a card's text with no pixel math.\n\nReach for explicit `x`\u002F`y` only for decoration (a stray mark, a background\nshape) or a deliberate free placement that has nothing to do with reading\norder. Anything a viewer reads — a title, a kicker, a caption, a row of\nstats — goes in a container.\n\n**Title block** — kicker, title, sub, stacked with a shared rhythm:\n\n```js\n\u002F\u002F fragment\nconst kicker = s.kicker(\"LAUNCHING TODAY\");\nconst title = s.text(\"Round-ups\", { size: 120, weight: 700 });\nconst sub = s.text(\"on every card\", { size: 40 });\ns.stack([kicker, title, sub], { gap: 20 });\n```\n\n**Lockup** — a mark beside its wordmark, one row:\n\n```js\n\u002F\u002F fragment\nconst mark = s.image(logoAssetId);\nconst wordmark = s.text(\"Northwind\", { size: 64, weight: 700 });\ns.row([mark, wordmark], { gap: 16, align: \"center\" });\n```\n\n**Card** — a padded column over an absolute plate:\n\n```js\n\u002F\u002F fragment\nconst plate = s.rect(1, 1, \"#0c2a21\", { r: 24, absolute: true });\nconst heading = s.text(\"Plan Pro\", { size: 48, weight: 700 });\nconst price = s.text(\"$12\u002Fmo\", { size: 32 });\ns.stack([plate, heading, price], { gap: 8, padding: 24 });\n```\n\n**List \u002F grid** — a row that wraps into a grid when it runs out of width:\n\n```js\n\u002F\u002F fragment\nconst chips = [\"Round-ups\", \"Instant transfers\", \"No fees\"].map((t) => s.pill(t));\ns.row(chips, { gap: 12, wrap: true });\n```\n\n## Grid\n\nTwelve columns for landscape, six for vertical. For a 1920-wide frame with a\n100px outer margin and a 24px gutter:\n\n```\ncontent = 1920 − 2×100 = 1720\ncolumn  = (1720 − 11×24) \u002F 12 = 121.3px\n```\n\nSpace everything in multiples of 8. `fontSizePx` for body copy sits near 2.5–4%\nof frame height; below 2.5% `validate_timeline` reports `text_illegible`.\nChoose the family, weights, and title scale from `motion-direction` § Typography\nbefore placing text on this grid.\n\n## One focal point\n\nPlace the subject on a thirds power point, not dead centre, unless the piece\nuses symmetry deliberately. The intersections:\n\n| Sequence | Power points (px) |\n|---|---|\n| 1920×1080 | x 640 \u002F 1280, y 360 \u002F 720 |\n| 1080×1920 | x 360 \u002F 720, y 640 \u002F 1280 |\n| 1080×1080 | x 360 \u002F 720, y 360 \u002F 720 |\n\nRank the eye's path by **size, then contrast, then colour, then position**. One\nprimary per frame; everything else supports. Leave negative space around it —\na frame filled edge to edge reads as cheap and gives the motion nowhere to go.\n\n## Depth: planes and track order\n\nSplit the picture into background, midground and foreground. Track index\ncontrols order **between** tracks: lowest index draws on top. Within one\ntrack, clips sort by `transform.depthPx`: farther values draw first and\nnearer, positive values draw later, in front. The same depth controls\nperspective under the sequence's `camera2d`. Moving a clip to another track\nchanges its overlap order but not its camera distance.\n\n| Plane | Suggested track index | Clip depth | Motion and look |\n|---|---|---|---|\n| Foreground | lowest | Nearer, positive | The move the eye follows; sharp unless focus shifts |\n| Midground | middle | Around zero | The subject and contrast peak |\n| Background | highest | Farther, negative | Small drift and lower contrast |\n\nFor flat parallax, run three speeds over one window. One `kenBurns` per layer\nwith `zoom: 0` and a different drift is a `fullClip` preset, so all three\ncover the same span. Put the three on clips of the same length: the drift is\nspread over the clip, so its speed scales with clip length:\n\n```json\n{\"role\": \"loop\", \"preset\": \"kenBurns\",\n \"params\": {\"zoom\": 0, \"driftX\": -0.02, \"driftY\": 0}}\n```\n\nBackground −0.02, midground −0.06, foreground −0.12. One layer at one speed is\nnot parallax, it is the template moving.\n\nFor a shared 2.5D move, give the layers distinct `transform.depthPx` values and\nkeyframe the sequence's `camera2d`. Its absolute `timeMs` keyframes interpolate\ncamera position and depth linearly. `focusDepthPx` and `aperturePx` control depth of\nfield. Use one camera move to connect the planes, then inspect a middle frame\nfor unwanted scale or blur. Set track indexes for overlap between tracks and\nclip depth for order within a track.\n\nFor a tilted card wall, parent the cards to a group with the wall's 2D\n`rotation`, `rotationX`, and `perspective`. Give each card its own `depthPx`\nand scroll curve, then use `camera2d` for the shared push and depth blur.\nThis is a flat card projection: preview the wall at several frames to check\nthat spacing, overlap, and blur read as one assembly.\n\nFor a move with a shape of its own, write a `custom` curve instead and mind the\nrole: an `in` window holds its `t=0` values before it and contributes nothing\nafter, an `out` window holds its `t=1` values after it, and a `loop` repeats its\ncycle — so a one-way travel that must stay where it landed is an `out`, and a\n`loop` curve has to return to the value it started on or it jumps every cycle.\n\nParent the assembly to a group when the layers must hold their relationship\nthrough a move: `add_group`, then animate the group. Children keep their own\ntracks, so grouping never changes what covers what.\nFor a transition on the assembled picture, set `transitionIn` on the group. The\nchildren stay parented on their own tracks while the transition acts once on\ntheir precomposed surface.\n\nShape geometry is rasterized inside a frame-sized source surface before the\nclip is placed. Keep normalized `shapeStyle` geometry inside that surface,\nthen move the clip or parent group with `transform.position` or custom\n`offsetX`\u002F`offsetY` curves for scrolling walls and flying cards. Geometry\nauthored outside the source surface clips before placement.\n\n## Camera moves\n\nChoose between a sequence camera move and a local clip move. `camera2d`\nkeyframes move the view across all depth planes. A layer animation moves one\nclip, and its `anchor` decides where scale or rotation pivots.\n\n| Move | How | Timing |\n|---|---|---|\n| Push \u002F pull | `kenBurns` with `zoom` 0.08–0.2 and `direction` in or out | 1500–4000ms, `easeInOut` |\n| Drift | `kenBurns` `driftX` \u002F `driftY`, ±0.02–0.08 | Same window as the zoom |\n| Pan | `custom` curve on `offsetX` across the clip | 800–2000ms, `easeInOut` |\n| Whip | `offsetX` over 150–250ms with a `blur` effect at 12–24px | On a cut, never mid-shot |\n| Follow | `offsetX` and `offsetY` tracking the subject, subject held off-centre and leading | Whatever the action takes |\n\n`camera2d` is a document field. `set_clip_params` changes a clip's `transform`\nbut refuses `camera2d`, `layout`, `repeater`, `motionBlur`, and other new document\nfields. Read the full document with `get_timeline`, change the relevant fields,\nthen write the full document with `set_timeline_document`. Keep its tracks,\nclips, markers, tempo, setup, media tracks, template identity, and other fields\nin the write. `timeline-edit-ops` carries the\nop contract.\n\n`kenBurns` is `fullClip`: it ignores `durationMs` and `delayMs` and runs across\nthe whole clip. To time a push to something shorter, write it as a `custom`\n`scale` curve instead.\n\nOne move per beat. A push plus a pan plus a rotate reads as chaos, and the three\nfold together into a direction nobody chose.\n\n## Enter and exit\n\nMotion that follows reading direction feels natural; reversing it reads as going\nback. Match `slide`'s `direction` to intent — it names the edge the element\narrives **from**:\n\n- New content: from the right, or from below.\n- Dismissing or undoing: reverse the entrance.\n- Drilling in: `pop`, or a scale curve rather than a travel.\n\nEnter and leave through the nearest edge. Keep `distance` at or below 0.35 or\nthe travel reads as a slide of the whole template — pair a longer move with a\nscale or opacity change on the same window. Same vocabulary on the cut:\n`set_transition` `direction` on `wipe`, `push` and `slide` also names the edge\nthe incoming clip arrives from.\n\n## Safe areas\n\nKeep titles, logos, faces and calls to action inside the margin. The numbers are\nfractions of the frame, so convert against `width` and `height`.\n\n| Aspect | Top | Bottom | Sides |\n|---|---|---|---|\n| 16:9 | 5% | 5% | 5% |\n| 9:16 | 12–14% | 18–20% | 6% |\n| 1:1 | 7% | 7% | 7% |\n| 4:5 | 8% | 8% | 6% |\n\nOn 1080×1920 that leaves the usable band roughly y 250 to y 1570. A caption clip\nalready sits at `bottomMarginFrac` 0.12 by default, which is inside the phone UI\nzone on some platforms — raise it rather than assuming.\n\n## One layout, several aspects\n\nMaster in the widest target with the focal point inside the centre band, then\n**restack** for vertical: move the headline above the subject, pull the subject\nup out of the bottom UI zone, and raise `fontSizePx` by about 20%. Cropping a\n16:9 layout to 9:16 clips whatever sat in the outer columns, which is usually the\nlogo and the call to action.\n\nBuilding in code, restacking is usually one option change on the container\nabove — flip `s.row` to `s.stack` (or its `align`\u002F`justify`), or move it to a\nnew `at`\u002F`anchor` — not a rebuild of every child's `x`\u002F`y`.\n\nOn the ops path, `layout: {display: \"flex\", flexDirection: \"row\" | \"column\",\ngap}` on a group clip keeps its real children (their `parentId` names the\ngroup) laid out with ordinary CSS flexbox — `justifyContent`\u002F`alignItems`\nposition them, `padding` insets the container. For a shape behind text, give\nthe shape `flexItem: {position: \"absolute\", inset: 0}` inside a padded flex\ncolumn holding the text: the shape's box resizes to the text's padded box,\nincluding an animated font size, with no manual pixel math. These relations\navoid redoing pixel math for each text change. A 16:9 document still needs a\nseparate placement pass for 9:16, with safe areas checked at that size.\n\n## Check the frame, not the document\n\n`preview_timeline_frame` composites the real picture — every track in order,\ntransforms and opacity applied, animations mid-flight. Read the layer list it\nreturns beside each frame: it names the stack top first with each layer's\n`z_index`, `opacity` and wipe progress, which answers most staging questions\nwithout looking at a pixel.\n\nPreview at the midpoint of every move, not at its ends. What frames catch and a\nvalidator cannot: a title that slid outside the frame, a scrim covering the face\nit was meant to sit beside, a lower third under the wrong element because two\ntrack indexes are the wrong way round, a layer that never draws because it is on\nan audio track.\n\n## Before you finish\n\n- One focal point, on a power point or deliberately centred.\n- Every element assigned a track for overlap order and a depth plane when the camera moves.\n- One camera move or eased layer move, 800–4000ms.\n- Entrances through the nearest edge, `distance` ≤ 0.35 or paired with a second\n  channel.\n- Critical content inside the safe margin for every target aspect, vertical\n  restacked rather than cropped.\n- Previewed at the midpoints, and the layer stack read.\n",{"name":5,"description":7},"main","c7226baa82daca42a19a5855c116544f97aa8bac","packages\u002Fsystem-skills\u002Fframe-composition",91,"https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool","https:\u002F\u002Fnodetool.ai\u002F","https:\u002F\u002Fgithub.com\u002Fnodetool-ai","No-code AI development","2026-11-15 03:27:49",{"source":8,"skillName":5,"commands":54,"prompt":64,"zipAvailable":65},[55,58,61],{"tool":56,"command":57},"npx","npx skills add nodetool-ai\u002Fnodetool --skill frame-composition",{"tool":59,"command":60},"bunx","bunx skills add nodetool-ai\u002Fnodetool --skill frame-composition",{"tool":62,"command":63},"pnpm","pnpm dlx skills add nodetool-ai\u002Fnodetool --skill frame-composition","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fframe-composition\u002FSKILL.md",true,[67],{"path":68,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":71,"sourceHtmlUrl":72},"references\u002Fdetails.md","REFERENCE","参考文档",10611,"https:\u002F\u002Fgithub.com\u002Fboisenoise\u002Fskills-collections\u002Fblob\u002Fmain\u002Fskills\u002Fantigravity-bb-methodology\u002Freferences\u002Fdetails.md",[74,83,117,147,162,178],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":78,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":82},"01M742MX3Y2ZYK884ZXTRFKB9P","humanizer","Rewrite AI-sounding text so it reads like the writer without changing what it says.\nUse when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line\nclosers, staged openers, forced triads, dashes everywhere, inflated claims, sales\nlanguage, stock AI words, bold labels, or filler.","MIT","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fhumanizer\u002FSKILL.md",[],[],"2026-11-11 15:11:14",{"skillNo":84,"name":36,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":98,"publishedAt":39},"01M6BY4D93R0Q9400ZHJ04DXP1","Use the NodeTool MCP server to build, validate, run, and debug visual AI workflows, and to work with assets, collections, and jobs. Use when the user mentions NodeTool or asks for a workflow graph, a node, or a media generation run on their NodeTool server.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fplugins\u002Fchatgpt\u002Fnodetool\u002Fskills\u002Fnodetool\u002FSKILL.md",[88,93],{"categoryNo":89,"categoryName":90,"taxonomyType":91,"taxonomyTypeDesc":92},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",{"categoryNo":94,"categoryName":95,"taxonomyType":96,"taxonomyTypeDesc":97},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[36,99,100,101,23,102,103,104,105,35,106,26,107,108,109,110,111,112,113,114,115,116],"MCP","Media","workflows","generation","media generation","with","visual","collections","node","work","debug","graph","validate","assets","run","server","mcp server","workflow",{"skillNo":118,"name":119,"nameZh":6,"description":120,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":121,"categories":122,"keywords":126,"publishedAt":39},"01M6BY4AEB05ZPBRVD078ATR2H","wan-2-6-prompting","Prompt Alibaba's Wan 2.6 video — text-to-video with a global style line plus timing-bracketed shots, and image-to-video that describes only what changes over time. Covers the shot-mode switch per provider, the prompt budget, negative prompts, motion intensity wording, audio input and prompt expansion. Use whenever the model id names a Wan 2.6 video route (alibaba\u002Fwan-2.6\u002Ftext-to-video, alibaba\u002Fwan-2.6\u002Fimage-to-video, wan-video\u002Fwan-2.6-t2v, wan-video\u002Fwan-2.6-i2v, wan-video\u002Fwan2.6-i2v-flash) on generate_video, animate_image or a TextToVideo node. Not for Wan 2.5 or 2.7, or the Wan 2.6 image routes.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fwan-2-6-prompting\u002FSKILL.md",[123],{"categoryNo":124,"categoryName":125,"taxonomyType":91,"taxonomyTypeDesc":92},"01JBSEEDCAT000000000000016","图像与多媒体",[127,128,129,130,131,132,133,107,134,135,136,137,138,139,140,141,142,143,144,145,146],"prompting","wan 2 6 prompting","line","expansion","input","image-to-video","text-to-video","route","only","mode","time","flash","Budget","generate","text","image","animate","model","Audio","Video",{"skillNo":148,"name":149,"nameZh":6,"description":150,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":151,"categories":152,"keywords":155,"publishedAt":39},"01M6BY47T4748ZG5ZMTAR97ZWZ","video-workflow","Build a reusable NodeTool video workflow for batches or repeated inputs. For a single finished piece, use a storyboard skill.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fvideo-workflow\u002FSKILL.md",[153,154],{"categoryNo":124,"categoryName":125,"taxonomyType":91,"taxonomyTypeDesc":92},{"categoryNo":89,"categoryName":90,"taxonomyType":91,"taxonomyTypeDesc":92},[116,156,146,157,158,159,160,35,161,36,26],"video workflow","reusable","storyboard","build","batches","single",{"skillNo":163,"name":164,"nameZh":6,"description":165,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":166,"categories":167,"keywords":169,"publishedAt":39},"01M6BY4587N2SYN315SXGEMZC3","video-clone","Analyze a reference video and recreate its shot structure with a new product or cast in NodeTool.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fvideo-clone\u002FSKILL.md",[168],{"categoryNo":124,"categoryName":125,"taxonomyType":91,"taxonomyTypeDesc":92},[170,171,146,104,172,173,36,174,175,176,177],"clone","video clone","cast","new","analyze","product","structure","reference",{"skillNo":179,"name":180,"nameZh":6,"description":181,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":182,"categories":183,"keywords":185,"publishedAt":39},"01M6BY4257C2KQYYMD9EK6DK8E","video-audio-continuity","Keep sound continuous across a multi-scene piece cut from generated video — why one clip per scene hard-cuts the audio at every boundary, when to write all the scenes into a single generation instead, the sound-design vocabulary that survives a provider's copyright filter, and how to check the length, the aspect ratio and the mix of what came back. Use when a piece has more than one scene and the video model writes its own audio (Seedance 2, Veo 3, MiniMax H3, Kling 2.6 and later), or when a cut's sound drops out at the shot changes.","https:\u002F\u002Fgithub.com\u002Fnodetool-ai\u002Fnodetool\u002Fblob\u002Fmain\u002Fpackages\u002Fsystem-skills\u002Fvideo-audio-continuity\u002FSKILL.md",[184],{"categoryNo":124,"categoryName":125,"taxonomyType":91,"taxonomyTypeDesc":92},[186,146,145,187,188,189,190,191,192,193,194,195,196,23,197,198,144,199,200,201,202],"video audio continuity","continuity","write","why","back","veo 3","at","cut","later","across","filter","generated","how to","mix","design","boundary","how"]