[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6D0DA8R4Y0WH92V9JA03E7B":3,"skill-same:01M6D0DA8R4Y0WH92V9JA03E7B":63},{"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":26,"linkOnly":27,"markdownContent":28,"markdownBody":29,"frontMatter":30,"refName":35,"commitSha":36,"skillDirectory":37,"forksCount":38,"repositoryHtmlUrl":39,"repositoryHomepageUrl":40,"ownerHtmlUrl":41,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":42,"install":43,"resources":56},"01M6D0DA8R4Y0WH92V9JA03E7B","designer_guide",null,"Guidance for creating and editing designs using the designer tool.","siddsachar\u002Frow-bot","siddsachar","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F50956096?v=4","01M6D0AG8BTX8VNFY5N3F2Q36T","Siddharth Sachar","01M6D0AGD3C7BZ6NAW9PWVKM3A",1687,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fdesigner_guide\u002FSKILL.md",[],[19,20,21,22,23,24,25],"designer","guide","designer guide","tool","editing","guidance","for","2026-11-10 10:12:30",false,"---\nname: designer_guide\ndisplay_name: Designer Guide\nicon: \"?\"\ndescription: Guidance for creating and editing designs using the designer tool.\ntools:\n  - designer\ntags: []\n---\nDESIGNER TOOL:\n- You have designer tools that create\u002Fedit multi-page visual designs.\n- Each project has a **mode**, a **canvas aspect ratio**, multi-page content, and an optional brand config.\n- All designs are rendered as HTML\u002FCSS in a sandboxed iframe preview.\n\n## Project modes (pick one — the mode dictates layout, canvas, and tool surface)\n\n| mode | What it produces | page_kind | Default aspect | Canvas behaviour |\n|---|---|---|---|---|\n| `deck` | Slide deck (16:9 \u002F 4:3 \u002F 1:1 \u002F 9:16) | slide | 16:9 (1920×1080) | **Fixed** body — `width:Wpx; height:Hpx; overflow:hidden`. Each page is a single slide. |\n| `document` | A4 \u002F Letter document pages | slide | A4 (794×1123) | **Fixed** page body — same fixed-frame rules as deck. |\n| `landing` | Tall scrollable landing page | screen | landing (1440×3200 hint) | **Tall scrollable**. Use `html,body{min-height:100vh;overflow-x:hidden}` and a `.page{max-width:1440px;margin:0 auto}` wrapper. NEVER set `body{overflow:hidden}` or fixed pixel `height`. Use responsive units (`vw`, `clamp()`, `%`). |\n| `app_mockup` | Phone or desktop UI prototype | screen | phone (390×844) | **One device viewport per page\u002Froute**. `html,body{width:Wpx;height:Hpx;overflow:hidden}`. In-content scrolling goes inside an inner `.screen-body{overflow-y:auto}`. |\n| `storyboard` | Cinematic shot-by-shot board | shot | 16:9 (1920×1080) | Fixed-frame like deck. Each page = one shot with beat \u002F caption \u002F camera direction. |\n\nThe user picks mode in the New Design dialog. You can read it with `designer_get_project` and change it with `designer_set_mode`. Switching modes does NOT auto-resize the canvas — call `designer_resize_project` after `designer_set_mode` if you also want a different aspect.\n\n## Core page operations\n- `designer_set_pages`: Replace ALL pages. Use for new projects or full reworks. Input: list of `{html, title, notes}`.\n- `designer_update_page`: Update a single page HTML. Input: `index`, `html`, optional `title`.\n- `designer_add_page`: Insert a new page. Use `index=-1` to append.\n- `designer_delete_page`: Remove a page by index. DESTRUCTIVE — requires user approval.\n- `designer_move_page`: Reorder a page. Input: `from_index`, `to_index`.\n- `designer_get_project`: Read project summary (mode, page titles, brand, dimensions). **Call before any multi-page edit.**\n- `designer_get_page_html`: Read the full stored HTML for a single page. Call before full-page rewrites so you preserve existing assets and layout.\n- `designer_get_reference`: Read a saved project reference by id or filename so you can reuse prior attachments without asking again.\n\n## Interactive modes — `landing` and `app_mockup`\nFor these modes, **never write raw `\u003Cscript>` or `onclick=`**. The preview is sandboxed and a thin runtime bridge wires interactivity from declarative attributes.\n\n### Declarative attributes the runtime understands\n- `data-row-bot-route=\"\u003Croute_id>\"` → marks a screen\u002Froute section. The runtime shows one route at a time. The outer route container also gets `data-row-bot-route-host=\"1\"` automatically when emitted via `designer_add_screen`.\n- `data-row-bot-action=\"navigate:\u003Croute_id>\"` → click handler that switches to the named route.\n- `data-row-bot-action=\"toggle_state:\u003Ckey>\"` → click handler that flips a state key. The runtime sets `data-row-bot-state-\u003Ckey>=\"on|off\"` on `\u003Chtml>`; style with `[data-row-bot-state-\u003Ckey>=\"on\"]` selectors.\n- `data-row-bot-action=\"play_media:\u003Casset_id>\"` → click handler that plays a video\u002Faudio asset.\n- `data-row-bot-transition=\"fade|slide_left|slide_up|none\"` (optional, on the navigation source).\n\n### Tools for interactive projects\n- `designer_add_screen`: Add a new screen\u002Froute. Input: `title`, optional `route_id` (auto-slugified from title), optional `html` (branded blank if empty), optional `copy_from` (page index to duplicate).\n- `designer_link_screens`: Wire a click on one element to navigate to another screen. Input: `source_route`, `selector` (CSS or `data-row-bot-element-id`), `target_route`, optional transition.\n- `designer_set_interaction`: Generic — attach `navigate` \u002F `toggle_state` \u002F `play_media` to any selector. Input: `source_route`, `selector`, `action`, `target`, optional `event`, optional `transition`.\n- `designer_reorder_routes`: Reorder the route list (which is the page list for screen-mode projects).\n- `designer_preview_screen`: Render a single route in isolation for review.\n- `designer_set_mode`: Switch the project's mode (and tool surface).\n\n### Authoring rules per interactive mode\n- **`landing`**: usually ONE page that scrolls vertically. Sections (`\u003Csection id=\"...\">`) for hero \u002F features \u002F pricing \u002F FAQ \u002F CTA \u002F footer. Use anchor links (`\u003Ca href=\"#features\">`) for in-page jumps; reserve `data-row-bot-action=\"navigate:...\"` for cross-route flows in multi-page landings.\n- **`app_mockup`**: ONE route per logical screen (Home, Detail, Settings, etc.). Each page's HTML is the entire device viewport. Use `data-row-bot-action=\"navigate:\u003Croute_id>\"` on tappable rows, tab bars, and back buttons. Use `data-row-bot-transition=\"slide_left\"` when going deeper, `slide_right` when going back. Wrap scrollable content inside an inner `.screen-body{overflow-y:auto}`.\n\n## Media (images, video)\n- `designer_generate_image`: Generate an AI image from a text prompt and embed it in a page. Input: `prompt`, optional `page_index` (-1=active), `position` (top\u002Fbottom), `width`, `height`, `size`. When a page contains a `data-row-bot-shot-visual` or `data-row-bot-image-slot` placeholder (storyboards, typed deck slots), the generated image automatically replaces the placeholder and fills the slot — no extra positioning call needed for the first generation.\n- `designer_insert_image`: Insert an attached or local image file. Same slot-replacement behavior as generate.\n- `designer_generate_video`: Generate an MP4 from a text prompt (or image-to-video). Input: `prompt`, optional `page_index`, `position`, `width`, `aspect_ratio`. Embedded as `\u003Cvideo autoplay loop muted playsinline>` with a poster.\n- `designer_insert_video`: Insert an attached \u002F local video file (mp4\u002Fwebm\u002Fmov). Same autoplay\u002Floop\u002Fmuted defaults.\n- `designer_move_image`, `designer_replace_image`, `designer_move_element`, `designer_duplicate_element`, `designer_restyle_element`: targeted DOM-level edits.\n- `designer_remove_image`: Remove an inserted image, chart, or video from a page WITHOUT deleting the page. Input: `image_ref` (asset ID or label), optional `page_index` (-1=active). Use this — NOT `designer_delete_page` and NOT `designer_replace_image` — when the user says \"remove the picture\", \"delete the image from shot 2\", \"clear the visual\", etc. Shot-visual \u002F image-slot placeholders revert to their dashed preview automatically.\n- When reusing a known project asset inside hand-written HTML, use `src=\"asset:\u002F\u002FASSET_ID\"` and keep `data-asset-id=\"ASSET_ID\"` on the `\u003Cimg>` \u002F `\u003Cvideo>`. Never invent placeholder tokens like `__ASSET_...__` or `__REMOVE__`.\n\n## AI content & polish\n- `designer_refine_text`: Refine a text element. Input: `page_index`, `tag` (e.g. 'h1', 'p'), `old_text` (exact text), `action` (shorten\u002Fexpand\u002Fprofessional\u002Fcasual\u002Fpersuasive\u002Fsimplify\u002Fbullets\u002Fparagraph\u002Fcustom), optional `custom_instruction`.\n- `designer_add_chart`: Add a chart. Input: `chart_type` (bar\u002Fline\u002Fpie\u002Fscatter\u002Fdonut\u002Fhistogram\u002Fbox\u002Farea\u002Fheatmap), `data_csv` (inline CSV with header), optional `title`, `page_index`, `position`.\n- `designer_insert_component`: Insert a curated reusable block (hero callout, stats band, testimonial, pricing cards, timeline section).\n- `designer_critique_page`: Review for hierarchy, overflow, contrast, readability, spacing.\n- `designer_apply_repairs`: Apply safe deterministic fixes for selected critique categories.\n- `designer_brand_lint`: Read-only scan for contrast issues, off-palette colors, non-brand fonts, missing alt text, logo safe-zone overlaps. Input: optional `page_index` (-1=all). Returns structured JSON.\n- `designer_generate_notes`: Generate speaker notes for one page (deck\u002Fdocument modes).\n\n## Brand & canvas\n- `designer_set_brand`: Update brand colors\u002Ffonts and logo placement. Includes `logo_mode`, `logo_scope`, `logo_position`, `logo_max_height`, `logo_padding`.\n- `designer_resize_project`: Resize the canvas using a built-in preset or explicit aspect ratio. Available aspects: `16:9`, `4:3`, `1:1`, `9:16`, `A4`, `letter`, `landing` (1440×3200), `phone` (390×844), `desktop` (1440×900). Use after `designer_set_mode` if changing both.\n\n## Export & publish\n- `designer_export`: Input: `format` (`pdf`\u002F`html`\u002F`png`\u002F`pptx`), optional page range. PDF\u002FPNG\u002FPPTX work best for `deck` and `document`. Single-file `html` works for any mode.\n- `designer_publish_link`: Publish a self-contained shareable HTML deck link. For `landing` and `app_mockup`, this produces an interactive bundle (includes the runtime bridge + transitions CSS) so navigation and toggles work in the published page.\n\n## HTML rules (all modes)\n- Each page MUST be a complete self-contained HTML document with inline `\u003Cstyle>`.\n- Use CSS variables for brand: `--primary`, `--secondary`, `--accent`, `--bg`, `--text`, `--heading-font`, `--body-font`.\n- Canvas size + mode are provided in the system prompt — follow the canvas rules table above.\n- Include Google Fonts `\u003Clink>` if using custom fonts.\n- Modern CSS only: flexbox, grid, gradients, box-shadow. No external frameworks.\n- For placeholder images: colored SVG shapes or gradient divs, never external URLs.\n- **No `\u003Cscript>` and no inline `onclick`** — the iframe is sandboxed. Use `data-row-bot-action` for interactive modes; deck\u002Fdocument\u002Fstoryboard pages have no interactivity.\n\n## Content budgets per mode (anti-clipping)\n\nFixed-slide modes (`deck`, `document`, `storyboard`) clip with `overflow:hidden` — anything past the canvas height is lost. Stick to these per-page budgets:\n\n- **`deck` (16:9 @ 1920×1080)**: ONE heading + EITHER (a) one paragraph ≤45 words, (b) up to 3 cards\u002Fcolumns with ≤25 words each, OR (c) one chart\u002Fimage + ≤2 bullets. Max 5 bullets. 64–96 px edge padding. Heading ≤4.5rem, body ≤1.4rem, line-height 1.3–1.5. Leave ≥48 px of bottom breathing room.\n- **`document` (A4 \u002F Letter)**: 130–160 words body copy per page (lead paragraph + 3–4 short sections with ~2 bullets each). Max ~5 top-level sections per page — split into more pages rather than cramming. Generous line-height (1.45–1.6), 16–24 px gap between sections, ≥32–48 px bottom padding.\n- **`storyboard` (1920×1080)**: each shot gets ONE eyebrow (≤3 words) + ONE heading (≤6 words) + ONE short paragraph (≤35 words) OR a pull-quote card + AT MOST two small metadata cards (shot type, camera move) OR one voiceover card — not both + ONE direction\u002Fnote line (≤30 words) + ONE footer strip (timeline OR shot counter). Pick 3–4 visible blocks max. Push extra detail to page notes via `designer_generate_notes` instead of the visible frame. Right column uses `display:flex; flex-direction:column; gap:20–28px; padding:64px 56px;` with ≥48 px bottom space.\n\nLanding and app_mockup modes flow differently — landing pages scroll vertically, app mockups fit one device viewport per route — but the same principle applies: count blocks before you write HTML, and split rather than cram.\n\n## Authoring guardrails (all modes)\n\n- **No decorative overlap on text.** Never place absolutely-positioned decorative CSS art (blobs, hand-drawn shapes, mascots, chef figures, etc.) on top of a heading or body paragraph. If a hero has a headline, either (a) give the illustration its own column\u002Frow, (b) put the illustration behind the text with low opacity and `z-index:0` AND ensure the text has a readable background or `text-shadow`, or (c) omit the illustration. When in doubt, use a real AI image via a typed image slot instead of CSS shapes.\n- **Button rows stay horizontal.** When emitting two buttons side-by-side (Back + primary, Cancel + Confirm, etc.), use `display:flex; gap:12–24px;` and give the primary button `flex:1`. Never let them stack vertically or touch. Secondary \u002F ghost buttons must be visually distinct (transparent or outlined) from the primary action — never two identical filled pills.\n- **Typed image slots over overlays.** When authoring a page that will later receive an AI image, mark the target container with `data-row-bot-image-slot=\"NAME\"` (e.g. `\u003Cdiv data-row-bot-image-slot=\"hero\" style=\"width:100%;aspect-ratio:16\u002F9;\">\u003C\u002Fdiv>`). `designer_generate_image` \u002F `designer_insert_image` will fill the first matching slot, sized to cover. Alternatively pass `position=\"replace:.my-class\"` or `position=\"replace:#my-id\"` to target a specific container. Never emit an absolute-positioned image overlay unless the user explicitly asked for a floating element.\n\n## Post-critique repair loop (mandatory after major rewrites)\n\nAfter any full-page rewrite, full-project rebuild, mode switch, canvas resize, or multi-page update:\n\n1. Call `designer_critique_page(page_index=\u003Cn>)` on each changed page.\n2. If any finding category is `overflow`, either:\n   - call `designer_apply_repairs(page_index=\u003Cn>, categories=[\"overflow\"])`, OR\n   - trim content and call `designer_update_page` again, OR\n   - split to a new page via `designer_add_page`.\n3. Never ship a page with an unresolved `overflow` finding.\n\nFor broader sweeps, pass multiple categories: `designer_apply_repairs(categories=[\"spacing\",\"readability\",\"overflow\"])`.\n\n## Workflow patterns\n\n### Decks \u002F documents \u002F storyboards (static)\n- Creating: `designer_get_project` → `designer_set_pages` with all pages.\n- Editing one slide: `designer_update_page` with the affected index.\n- \"Make all pages darker\": `designer_get_project` → `designer_update_page` per page.\n- \"Add a pricing slide after slide 3\": `designer_add_page(index=3, ...)`.\n- \"Export as PDF\": `designer_export(format=\"pdf\")`.\n- Speaker notes: `designer_generate_notes` for the relevant page (don't rewrite the HTML).\n\n### Landing pages\n- One page is fine. Build vertically with `\u003Csection>` blocks: hero, features, pricing, FAQ, CTA, footer.\n- Use `clamp()` and `vw` units for typography. Use `max-width:1440px` on a `.page` wrapper.\n- For multi-page landing flows (e.g. a stepper), use `designer_add_screen` per step and `designer_link_screens` for the navigation.\n\n### App mockups\n- Plan routes first: list every screen the user mentions. Call `designer_set_pages` (or `designer_add_screen` per route) so each page has `data-row-bot-route=\"\u003Cid>\"` on its outer container.\n- Wire navigation explicitly with `designer_link_screens` or `designer_set_interaction(action=\"navigate\", ...)`. Don't use `\u003Ca href=\"#detail\">`.\n- For toggleable UI (dark mode, notifications, expanded panels), use `data-row-bot-action=\"toggle_state:\u003Ckey>\"` and style with `[data-row-bot-state-\u003Ckey>=\"on\"]` selectors on `\u003Chtml>`.\n- Wrap scrollable content inside an inner `.screen-body` — the outer `body` stays at the device viewport.\n\n### Mode change\n- \"Turn this deck into a landing page\": `designer_set_mode(\"landing\")` → `designer_resize_project(\"landing\")` → `designer_set_pages([...])` rewriting the content as a single tall page.\n\n### Canvas \u002F brand\n- Canvas changes: `designer_resize_project` before restyling when the user wants square, vertical, A4, Letter, phone, desktop, landing, or standard slide formats.\n- Brand changes: `designer_set_brand` first. It updates stored brand CSS and can switch the automatic logo overlay between all pages, first page only, or manual placeholder mode.\n\n### Sharing\n- Shareable deck\u002Flanding URL: `designer_publish_link` (handles interactive bundles automatically when mode is `landing` or `app_mockup`).\n- One-off file: `designer_export(format=\"html\")` for a single self-contained file.\n\n### AI imagery\n- \"Add a photo of mountains\": `designer_generate_image(prompt=\"cinematic photo of mountain landscape at sunrise\")` — or ask the user to attach a reference image and call `designer_insert_image`.\n- \"Generate an AI image of a futuristic city\": `designer_generate_image(prompt=\"futuristic city skyline\")`.\n- \"Add a 4-second product video\": `designer_generate_video(prompt=\"...\", aspect_ratio=\"16:9\")`.\n- \"Generate a still image for every shot\": call `designer_generate_image` once per page — shot-visual placeholders are filled automatically in the correct slot for each.\n- \"Remove the picture from shot 2\" \u002F \"delete the image on this page\": call `designer_remove_image(image_ref=\"\u003Casset-id or label>\", page_index=\u003Cn>)`. Do NOT use `designer_delete_page` (that removes the whole page) and do NOT pass a fake sentinel to `designer_replace_image`.\n\n### Components & critiques\n- \"Add a metrics strip near the top\": `designer_insert_component(component_name=\"stats_band\", page_index=-1, position=\"top\")`.\n- \"Review the current slide and fix what feels cramped\": `designer_critique_page(page_index=-1)` → `designer_apply_repairs(page_index=-1, categories=[\"spacing\",\"readability\",\"overflow\"])`.\n- \"Make the heading shorter\": `designer_refine_text(page_index=0, tag=\"h1\", old_text=\"...\", action=\"shorten\")`.\n- \"Add a bar chart of Q1 revenue\": `designer_add_chart(chart_type=\"bar\", data_csv=\"Quarter,Revenue\\nQ1,120\\nQ2,150\\n...\")`.\n- \"Show me what you changed\" \u002F \"review last turn\": the UI surfaces a review dialog with a page-by-page mutation diff after every agent turn — you don't need to call it, but do reference it when the user asks what you altered.\n\n## IMPORTANT\n- Refs from `designer_get_project` can go stale. Re-read if unsure.\n- Always explain what you changed in your text response after tool calls.\n- Maintain visual consistency across pages.\n- Keep HTML compact — avoid unnecessary nesting or unused CSS.\n- AI images: descriptive prompts. Specify style (photo, illustration, etc.). There is no stock-image tool — use `designer_generate_image` or `designer_insert_image` (attached\u002Flocal file) only.\n- For interactive modes, **never** emit `\u003Cscript>`, inline event handlers, or rely on `\u003Ca href>` for navigation between routes — use `data-row-bot-action`.\n","DESIGNER TOOL:\n- You have designer tools that create\u002Fedit multi-page visual designs.\n- Each project has a **mode**, a **canvas aspect ratio**, multi-page content, and an optional brand config.\n- All designs are rendered as HTML\u002FCSS in a sandboxed iframe preview.\n\n## Project modes (pick one — the mode dictates layout, canvas, and tool surface)\n\n| mode | What it produces | page_kind | Default aspect | Canvas behaviour |\n|---|---|---|---|---|\n| `deck` | Slide deck (16:9 \u002F 4:3 \u002F 1:1 \u002F 9:16) | slide | 16:9 (1920×1080) | **Fixed** body — `width:Wpx; height:Hpx; overflow:hidden`. Each page is a single slide. |\n| `document` | A4 \u002F Letter document pages | slide | A4 (794×1123) | **Fixed** page body — same fixed-frame rules as deck. |\n| `landing` | Tall scrollable landing page | screen | landing (1440×3200 hint) | **Tall scrollable**. Use `html,body{min-height:100vh;overflow-x:hidden}` and a `.page{max-width:1440px;margin:0 auto}` wrapper. NEVER set `body{overflow:hidden}` or fixed pixel `height`. Use responsive units (`vw`, `clamp()`, `%`). |\n| `app_mockup` | Phone or desktop UI prototype | screen | phone (390×844) | **One device viewport per page\u002Froute**. `html,body{width:Wpx;height:Hpx;overflow:hidden}`. In-content scrolling goes inside an inner `.screen-body{overflow-y:auto}`. |\n| `storyboard` | Cinematic shot-by-shot board | shot | 16:9 (1920×1080) | Fixed-frame like deck. Each page = one shot with beat \u002F caption \u002F camera direction. |\n\nThe user picks mode in the New Design dialog. You can read it with `designer_get_project` and change it with `designer_set_mode`. Switching modes does NOT auto-resize the canvas — call `designer_resize_project` after `designer_set_mode` if you also want a different aspect.\n\n## Core page operations\n- `designer_set_pages`: Replace ALL pages. Use for new projects or full reworks. Input: list of `{html, title, notes}`.\n- `designer_update_page`: Update a single page HTML. Input: `index`, `html`, optional `title`.\n- `designer_add_page`: Insert a new page. Use `index=-1` to append.\n- `designer_delete_page`: Remove a page by index. DESTRUCTIVE — requires user approval.\n- `designer_move_page`: Reorder a page. Input: `from_index`, `to_index`.\n- `designer_get_project`: Read project summary (mode, page titles, brand, dimensions). **Call before any multi-page edit.**\n- `designer_get_page_html`: Read the full stored HTML for a single page. Call before full-page rewrites so you preserve existing assets and layout.\n- `designer_get_reference`: Read a saved project reference by id or filename so you can reuse prior attachments without asking again.\n\n## Interactive modes — `landing` and `app_mockup`\nFor these modes, **never write raw `\u003Cscript>` or `onclick=`**. The preview is sandboxed and a thin runtime bridge wires interactivity from declarative attributes.\n\n### Declarative attributes the runtime understands\n- `data-row-bot-route=\"\u003Croute_id>\"` → marks a screen\u002Froute section. The runtime shows one route at a time. The outer route container also gets `data-row-bot-route-host=\"1\"` automatically when emitted via `designer_add_screen`.\n- `data-row-bot-action=\"navigate:\u003Croute_id>\"` → click handler that switches to the named route.\n- `data-row-bot-action=\"toggle_state:\u003Ckey>\"` → click handler that flips a state key. The runtime sets `data-row-bot-state-\u003Ckey>=\"on|off\"` on `\u003Chtml>`; style with `[data-row-bot-state-\u003Ckey>=\"on\"]` selectors.\n- `data-row-bot-action=\"play_media:\u003Casset_id>\"` → click handler that plays a video\u002Faudio asset.\n- `data-row-bot-transition=\"fade|slide_left|slide_up|none\"` (optional, on the navigation source).\n\n### Tools for interactive projects\n- `designer_add_screen`: Add a new screen\u002Froute. Input: `title`, optional `route_id` (auto-slugified from title), optional `html` (branded blank if empty), optional `copy_from` (page index to duplicate).\n- `designer_link_screens`: Wire a click on one element to navigate to another screen. Input: `source_route`, `selector` (CSS or `data-row-bot-element-id`), `target_route`, optional transition.\n- `designer_set_interaction`: Generic — attach `navigate` \u002F `toggle_state` \u002F `play_media` to any selector. Input: `source_route`, `selector`, `action`, `target`, optional `event`, optional `transition`.\n- `designer_reorder_routes`: Reorder the route list (which is the page list for screen-mode projects).\n- `designer_preview_screen`: Render a single route in isolation for review.\n- `designer_set_mode`: Switch the project's mode (and tool surface).\n\n### Authoring rules per interactive mode\n- **`landing`**: usually ONE page that scrolls vertically. Sections (`\u003Csection id=\"...\">`) for hero \u002F features \u002F pricing \u002F FAQ \u002F CTA \u002F footer. Use anchor links (`\u003Ca href=\"#features\">`) for in-page jumps; reserve `data-row-bot-action=\"navigate:...\"` for cross-route flows in multi-page landings.\n- **`app_mockup`**: ONE route per logical screen (Home, Detail, Settings, etc.). Each page's HTML is the entire device viewport. Use `data-row-bot-action=\"navigate:\u003Croute_id>\"` on tappable rows, tab bars, and back buttons. Use `data-row-bot-transition=\"slide_left\"` when going deeper, `slide_right` when going back. Wrap scrollable content inside an inner `.screen-body{overflow-y:auto}`.\n\n## Media (images, video)\n- `designer_generate_image`: Generate an AI image from a text prompt and embed it in a page. Input: `prompt`, optional `page_index` (-1=active), `position` (top\u002Fbottom), `width`, `height`, `size`. When a page contains a `data-row-bot-shot-visual` or `data-row-bot-image-slot` placeholder (storyboards, typed deck slots), the generated image automatically replaces the placeholder and fills the slot — no extra positioning call needed for the first generation.\n- `designer_insert_image`: Insert an attached or local image file. Same slot-replacement behavior as generate.\n- `designer_generate_video`: Generate an MP4 from a text prompt (or image-to-video). Input: `prompt`, optional `page_index`, `position`, `width`, `aspect_ratio`. Embedded as `\u003Cvideo autoplay loop muted playsinline>` with a poster.\n- `designer_insert_video`: Insert an attached \u002F local video file (mp4\u002Fwebm\u002Fmov). Same autoplay\u002Floop\u002Fmuted defaults.\n- `designer_move_image`, `designer_replace_image`, `designer_move_element`, `designer_duplicate_element`, `designer_restyle_element`: targeted DOM-level edits.\n- `designer_remove_image`: Remove an inserted image, chart, or video from a page WITHOUT deleting the page. Input: `image_ref` (asset ID or label), optional `page_index` (-1=active). Use this — NOT `designer_delete_page` and NOT `designer_replace_image` — when the user says \"remove the picture\", \"delete the image from shot 2\", \"clear the visual\", etc. Shot-visual \u002F image-slot placeholders revert to their dashed preview automatically.\n- When reusing a known project asset inside hand-written HTML, use `src=\"asset:\u002F\u002FASSET_ID\"` and keep `data-asset-id=\"ASSET_ID\"` on the `\u003Cimg>` \u002F `\u003Cvideo>`. Never invent placeholder tokens like `__ASSET_...__` or `__REMOVE__`.\n\n## AI content & polish\n- `designer_refine_text`: Refine a text element. Input: `page_index`, `tag` (e.g. 'h1', 'p'), `old_text` (exact text), `action` (shorten\u002Fexpand\u002Fprofessional\u002Fcasual\u002Fpersuasive\u002Fsimplify\u002Fbullets\u002Fparagraph\u002Fcustom), optional `custom_instruction`.\n- `designer_add_chart`: Add a chart. Input: `chart_type` (bar\u002Fline\u002Fpie\u002Fscatter\u002Fdonut\u002Fhistogram\u002Fbox\u002Farea\u002Fheatmap), `data_csv` (inline CSV with header), optional `title`, `page_index`, `position`.\n- `designer_insert_component`: Insert a curated reusable block (hero callout, stats band, testimonial, pricing cards, timeline section).\n- `designer_critique_page`: Review for hierarchy, overflow, contrast, readability, spacing.\n- `designer_apply_repairs`: Apply safe deterministic fixes for selected critique categories.\n- `designer_brand_lint`: Read-only scan for contrast issues, off-palette colors, non-brand fonts, missing alt text, logo safe-zone overlaps. Input: optional `page_index` (-1=all). Returns structured JSON.\n- `designer_generate_notes`: Generate speaker notes for one page (deck\u002Fdocument modes).\n\n## Brand & canvas\n- `designer_set_brand`: Update brand colors\u002Ffonts and logo placement. Includes `logo_mode`, `logo_scope`, `logo_position`, `logo_max_height`, `logo_padding`.\n- `designer_resize_project`: Resize the canvas using a built-in preset or explicit aspect ratio. Available aspects: `16:9`, `4:3`, `1:1`, `9:16`, `A4`, `letter`, `landing` (1440×3200), `phone` (390×844), `desktop` (1440×900). Use after `designer_set_mode` if changing both.\n\n## Export & publish\n- `designer_export`: Input: `format` (`pdf`\u002F`html`\u002F`png`\u002F`pptx`), optional page range. PDF\u002FPNG\u002FPPTX work best for `deck` and `document`. Single-file `html` works for any mode.\n- `designer_publish_link`: Publish a self-contained shareable HTML deck link. For `landing` and `app_mockup`, this produces an interactive bundle (includes the runtime bridge + transitions CSS) so navigation and toggles work in the published page.\n\n## HTML rules (all modes)\n- Each page MUST be a complete self-contained HTML document with inline `\u003Cstyle>`.\n- Use CSS variables for brand: `--primary`, `--secondary`, `--accent`, `--bg`, `--text`, `--heading-font`, `--body-font`.\n- Canvas size + mode are provided in the system prompt — follow the canvas rules table above.\n- Include Google Fonts `\u003Clink>` if using custom fonts.\n- Modern CSS only: flexbox, grid, gradients, box-shadow. No external frameworks.\n- For placeholder images: colored SVG shapes or gradient divs, never external URLs.\n- **No `\u003Cscript>` and no inline `onclick`** — the iframe is sandboxed. Use `data-row-bot-action` for interactive modes; deck\u002Fdocument\u002Fstoryboard pages have no interactivity.\n\n## Content budgets per mode (anti-clipping)\n\nFixed-slide modes (`deck`, `document`, `storyboard`) clip with `overflow:hidden` — anything past the canvas height is lost. Stick to these per-page budgets:\n\n- **`deck` (16:9 @ 1920×1080)**: ONE heading + EITHER (a) one paragraph ≤45 words, (b) up to 3 cards\u002Fcolumns with ≤25 words each, OR (c) one chart\u002Fimage + ≤2 bullets. Max 5 bullets. 64–96 px edge padding. Heading ≤4.5rem, body ≤1.4rem, line-height 1.3–1.5. Leave ≥48 px of bottom breathing room.\n- **`document` (A4 \u002F Letter)**: 130–160 words body copy per page (lead paragraph + 3–4 short sections with ~2 bullets each). Max ~5 top-level sections per page — split into more pages rather than cramming. Generous line-height (1.45–1.6), 16–24 px gap between sections, ≥32–48 px bottom padding.\n- **`storyboard` (1920×1080)**: each shot gets ONE eyebrow (≤3 words) + ONE heading (≤6 words) + ONE short paragraph (≤35 words) OR a pull-quote card + AT MOST two small metadata cards (shot type, camera move) OR one voiceover card — not both + ONE direction\u002Fnote line (≤30 words) + ONE footer strip (timeline OR shot counter). Pick 3–4 visible blocks max. Push extra detail to page notes via `designer_generate_notes` instead of the visible frame. Right column uses `display:flex; flex-direction:column; gap:20–28px; padding:64px 56px;` with ≥48 px bottom space.\n\nLanding and app_mockup modes flow differently — landing pages scroll vertically, app mockups fit one device viewport per route — but the same principle applies: count blocks before you write HTML, and split rather than cram.\n\n## Authoring guardrails (all modes)\n\n- **No decorative overlap on text.** Never place absolutely-positioned decorative CSS art (blobs, hand-drawn shapes, mascots, chef figures, etc.) on top of a heading or body paragraph. If a hero has a headline, either (a) give the illustration its own column\u002Frow, (b) put the illustration behind the text with low opacity and `z-index:0` AND ensure the text has a readable background or `text-shadow`, or (c) omit the illustration. When in doubt, use a real AI image via a typed image slot instead of CSS shapes.\n- **Button rows stay horizontal.** When emitting two buttons side-by-side (Back + primary, Cancel + Confirm, etc.), use `display:flex; gap:12–24px;` and give the primary button `flex:1`. Never let them stack vertically or touch. Secondary \u002F ghost buttons must be visually distinct (transparent or outlined) from the primary action — never two identical filled pills.\n- **Typed image slots over overlays.** When authoring a page that will later receive an AI image, mark the target container with `data-row-bot-image-slot=\"NAME\"` (e.g. `\u003Cdiv data-row-bot-image-slot=\"hero\" style=\"width:100%;aspect-ratio:16\u002F9;\">\u003C\u002Fdiv>`). `designer_generate_image` \u002F `designer_insert_image` will fill the first matching slot, sized to cover. Alternatively pass `position=\"replace:.my-class\"` or `position=\"replace:#my-id\"` to target a specific container. Never emit an absolute-positioned image overlay unless the user explicitly asked for a floating element.\n\n## Post-critique repair loop (mandatory after major rewrites)\n\nAfter any full-page rewrite, full-project rebuild, mode switch, canvas resize, or multi-page update:\n\n1. Call `designer_critique_page(page_index=\u003Cn>)` on each changed page.\n2. If any finding category is `overflow`, either:\n   - call `designer_apply_repairs(page_index=\u003Cn>, categories=[\"overflow\"])`, OR\n   - trim content and call `designer_update_page` again, OR\n   - split to a new page via `designer_add_page`.\n3. Never ship a page with an unresolved `overflow` finding.\n\nFor broader sweeps, pass multiple categories: `designer_apply_repairs(categories=[\"spacing\",\"readability\",\"overflow\"])`.\n\n## Workflow patterns\n\n### Decks \u002F documents \u002F storyboards (static)\n- Creating: `designer_get_project` → `designer_set_pages` with all pages.\n- Editing one slide: `designer_update_page` with the affected index.\n- \"Make all pages darker\": `designer_get_project` → `designer_update_page` per page.\n- \"Add a pricing slide after slide 3\": `designer_add_page(index=3, ...)`.\n- \"Export as PDF\": `designer_export(format=\"pdf\")`.\n- Speaker notes: `designer_generate_notes` for the relevant page (don't rewrite the HTML).\n\n### Landing pages\n- One page is fine. Build vertically with `\u003Csection>` blocks: hero, features, pricing, FAQ, CTA, footer.\n- Use `clamp()` and `vw` units for typography. Use `max-width:1440px` on a `.page` wrapper.\n- For multi-page landing flows (e.g. a stepper), use `designer_add_screen` per step and `designer_link_screens` for the navigation.\n\n### App mockups\n- Plan routes first: list every screen the user mentions. Call `designer_set_pages` (or `designer_add_screen` per route) so each page has `data-row-bot-route=\"\u003Cid>\"` on its outer container.\n- Wire navigation explicitly with `designer_link_screens` or `designer_set_interaction(action=\"navigate\", ...)`. Don't use `\u003Ca href=\"#detail\">`.\n- For toggleable UI (dark mode, notifications, expanded panels), use `data-row-bot-action=\"toggle_state:\u003Ckey>\"` and style with `[data-row-bot-state-\u003Ckey>=\"on\"]` selectors on `\u003Chtml>`.\n- Wrap scrollable content inside an inner `.screen-body` — the outer `body` stays at the device viewport.\n\n### Mode change\n- \"Turn this deck into a landing page\": `designer_set_mode(\"landing\")` → `designer_resize_project(\"landing\")` → `designer_set_pages([...])` rewriting the content as a single tall page.\n\n### Canvas \u002F brand\n- Canvas changes: `designer_resize_project` before restyling when the user wants square, vertical, A4, Letter, phone, desktop, landing, or standard slide formats.\n- Brand changes: `designer_set_brand` first. It updates stored brand CSS and can switch the automatic logo overlay between all pages, first page only, or manual placeholder mode.\n\n### Sharing\n- Shareable deck\u002Flanding URL: `designer_publish_link` (handles interactive bundles automatically when mode is `landing` or `app_mockup`).\n- One-off file: `designer_export(format=\"html\")` for a single self-contained file.\n\n### AI imagery\n- \"Add a photo of mountains\": `designer_generate_image(prompt=\"cinematic photo of mountain landscape at sunrise\")` — or ask the user to attach a reference image and call `designer_insert_image`.\n- \"Generate an AI image of a futuristic city\": `designer_generate_image(prompt=\"futuristic city skyline\")`.\n- \"Add a 4-second product video\": `designer_generate_video(prompt=\"...\", aspect_ratio=\"16:9\")`.\n- \"Generate a still image for every shot\": call `designer_generate_image` once per page — shot-visual placeholders are filled automatically in the correct slot for each.\n- \"Remove the picture from shot 2\" \u002F \"delete the image on this page\": call `designer_remove_image(image_ref=\"\u003Casset-id or label>\", page_index=\u003Cn>)`. Do NOT use `designer_delete_page` (that removes the whole page) and do NOT pass a fake sentinel to `designer_replace_image`.\n\n### Components & critiques\n- \"Add a metrics strip near the top\": `designer_insert_component(component_name=\"stats_band\", page_index=-1, position=\"top\")`.\n- \"Review the current slide and fix what feels cramped\": `designer_critique_page(page_index=-1)` → `designer_apply_repairs(page_index=-1, categories=[\"spacing\",\"readability\",\"overflow\"])`.\n- \"Make the heading shorter\": `designer_refine_text(page_index=0, tag=\"h1\", old_text=\"...\", action=\"shorten\")`.\n- \"Add a bar chart of Q1 revenue\": `designer_add_chart(chart_type=\"bar\", data_csv=\"Quarter,Revenue\\nQ1,120\\nQ2,150\\n...\")`.\n- \"Show me what you changed\" \u002F \"review last turn\": the UI surfaces a review dialog with a page-by-page mutation diff after every agent turn — you don't need to call it, but do reference it when the user asks what you altered.\n\n## IMPORTANT\n- Refs from `designer_get_project` can go stale. Re-read if unsure.\n- Always explain what you changed in your text response after tool calls.\n- Maintain visual consistency across pages.\n- Keep HTML compact — avoid unnecessary nesting or unused CSS.\n- AI images: descriptive prompts. Specify style (photo, illustration, etc.). There is no stock-image tool — use `designer_generate_image` or `designer_insert_image` (attached\u002Flocal file) only.\n- For interactive modes, **never** emit `\u003Cscript>`, inline event handlers, or rely on `\u003Ca href>` for navigation between routes — use `data-row-bot-action`.\n",{"icon":31,"name":5,"tags":32,"tools":33,"description":7,"display_name":34},"?",[],[19],"Designer Guide","main","aea3219498f535052aba82436fc9612f548a086e","tool_guides\u002Fdesigner_guide",204,"https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot","https:\u002F\u002Frow-bot.ai","https:\u002F\u002Fgithub.com\u002Fsiddsachar","2026-11-15 02:36:29",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add siddsachar\u002Frow-bot --skill designer_guide",{"tool":49,"command":50},"bunx","bunx skills add siddsachar\u002Frow-bot --skill designer_guide",{"tool":52,"command":53},"pnpm","pnpm dlx skills add siddsachar\u002Frow-bot --skill designer_guide","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fdesigner_guide\u002FSKILL.md",true,[57],{"path":58,"resourceType":59,"resourceTypeDesc":60,"sizeBytes":61,"sourceHtmlUrl":62},"evals\u002Fevals.json","REFERENCE","参考文档",2239,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fskills_lint.dart\u002Fblob\u002Fmain\u002Fthird_party\u002Fskill-repos\u002Fkevmoo-dash-skills\u002F.agents\u002Fskills\u002Fdart-long-lines\u002Fevals\u002Fevals.json",[64,83,108,119,129,144],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":26},"01M6D0EZEJNX7R2J8EXN3386BW","x_guide","Essential guidance for using X (Twitter) tools correctly — search syntax, time filtering, and engagement patterns.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fx_guide\u002FSKILL.md",[],[71,20,25,72,73,74,75,76,77,78,24,79,80,81,82],"x guide","bot","search","Tools","syntax","twitter","filtering","engagement","patterns","social","time","essential",{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":94,"publishedAt":26},"01M6D0EWESTPTF4B1JZFVFQ49Y","workflow_guide","How workflows run (scheduled chats), how to write and change them, hand work between them, and check what a run did.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fworkflow_guide\u002FSKILL.md",[89],{"categoryNo":90,"categoryName":91,"taxonomyType":92,"taxonomyTypeDesc":93},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[95,20,96,97,98,99,100,101,102,103,104,105,106,107],"workflow guide","workflow","write","task","how to","hand","scheduled","check","workflows","change","run","how","work",{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":26},"01M6D0ESS92MEV9V5EAHJ3N6VR","wiki_guide","Guidance for the Obsidian-compatible wiki vault.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fwiki_guide\u002FSKILL.md",[],[115,20,116,117,118,24,25],"wiki guide","wiki","obsidian","vault",{"skillNo":120,"name":121,"nameZh":6,"description":122,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":123,"categories":124,"keywords":125,"publishedAt":26},"01M6D0EQDFABHZ27HMDKYSX6V4","weather_guide","Guidance for using weather tools effectively.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fweather_guide\u002FSKILL.md",[],[126,20,127,74,128,24,25],"weather guide","weather","weather tools",{"skillNo":130,"name":131,"nameZh":6,"description":132,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":133,"categories":134,"keywords":135,"publishedAt":26},"01M6D0EMVZ8MZHAVSTMMTMBWN0","vision_guide","Guidance for using webcam, screen capture, and image analysis.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fvision_guide\u002FSKILL.md",[],[136,137,20,24,138,139,140,25,141,142,143],"vision guide","vision","analysis","screen","image analysis","capture","screen capture","image",{"skillNo":145,"name":146,"nameZh":6,"description":147,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":148,"categories":149,"keywords":153,"publishedAt":26},"01M6D0EJF8QMGWNNKD38BK24G8","video_guide","Guidance for generating videos and animating images.","https:\u002F\u002Fgithub.com\u002Fsiddsachar\u002Frow-bot\u002Fblob\u002Fmain\u002Ftool_guides\u002Fvideo_guide\u002FSKILL.md",[150],{"categoryNo":151,"categoryName":152,"taxonomyType":92,"taxonomyTypeDesc":93},"01JBSEEDCAT000000000000016","图像与多媒体",[20,154,155,24,156,25,157],"Video","video guide","images","videos"]