Zhongcao OOTD Lookbook Maker

--- name: "zhongcao-ootd-lookbook-maker" description: "Create a coordinated REDnote (Xiaohongshu) OOTD lookbook from outfit photos or a styling idea. Build a vertical 3:4 fashion carousel with a cover, full-look outfit image, styling-detail image, and lifestyle scene, then shape a ready-to-publish fashion recommendation post with title ideas, caption angles, and tags. Use this AI outfit image maker for Xiaohongshu outfit posts, OOTD photos, fashion lookbooks, clothing recommendation visuals, creator style diaries, and brand collaboration campaigns. Optionally it reads Xiaohongshu itself — the notes already running for the topic, one page of a note's top comments, and an account's own recent notes — so Xiaohongshu research, competitor note analysis and comment analysis rest on the platform instead of on guesswork." --- # Zhongcao OOTD Lookbook Maker Create an ordered REDnote (Xiaohongshu) OOTD lookbook that gives one outfit a complete visual story: a cover, full-look image, styling detail, and lifestyle scene. Pair the visual sequence with a practical post angle, title ideas, caption beats, and tags in the user's voice. ## Scope and routing Use this Skill for a coordinated outfit-led carousel, a fashion diary, a clothing recommendation post, or brand collaboration campaign where the creator wants multiple images to feel like one story. A typical request starts with an outfit photo, mirror selfie, flat lay, garment image, or a clear styling concept. For one isolated REDnote cover, route to `zhongcao-cover-maker`. For a market-specific on-model ecommerce visual from a confirmed wearable SKU, route to `product-on-model-locale-studio`. Keep the user's stated garment details, accessories, occasion, and visual references central whenever this Skill owns the lookbook. ## Inputs and defaults Reuse the conversation's outfit, occasion, target audience, style vocabulary, and references. A source outfit photo or a concrete styling idea is the minimum hard input. Ask only for a missing choice that materially changes the visual story: the outfit, the intended occasion, or the style direction. With an outfit photo, use it as the first ordered reference for every outfit-preserving image. With a styling idea, create an original lookbook from the user's described garments, palette, occasion, and mood. Default to a four-slide vertical `3:4` lookbook at `2K`, delivered as one coordinated sequence when the live model card accepts `count: 4` and `output_relationship: "sequence"`: 1. Cover — a clear outfit introduction with clean title space. 2. Full look — a readable head-to-toe view that makes the silhouette and layering easy to understand. 3. Detail — the material, accessory, color relationship, or styling move that makes the outfit memorable. 4. Lifestyle scene — an occasion-led final image that carries the look's mood. Keep `model: "auto"` and model-managed controls unless the user asks for a model, compatibility, or price decision. Read `beatra.models.list` for the selected `image_to_image`, `text_to_image`, or `image_edit` capability before fixing a model, canvas, control, count, output relationship, or current price. When the live card does not accept the coordinated four-image sequence, present its compatible routes and their maximum charge before the user selects any different paid plan. ## Golden path 1. Build one lookbook card: outfit must-keeps, occasion, audience, story angle, palette, light, setting, title-safe placement, and each slide's role. 2. Choose the route: `beatra.images.transform` for a source outfit photo; `beatra.images.generate` for a styling idea; `beatra.images.edit` for a focused revision to an accepted slide. 3. Draft the four prompts as one visual family. Keep source and later reference roles explicit, and create the post angle, title ideas, caption beats, and tags before paid execution. 4. Read the live model card, then show one confirmation containing the ordered four-slide sequence, all slide roles, the complete prompt, ordered references, canvas, count, output relationship, model behaviour, current maximum charge, and call count. 5. After approval, assign the coordinated sequence one stable opaque `client_request_id`, submit it once through the bundled client, and retain the returned task ID. 6. Poll the original task, review accessible outputs against the lookbook card, and deliver the images in post order with the caption plan and actual returned image and billing facts. Read [lookbook planning](references/lookbook-planning.md) to shape the visual story and [lookbook workflow](references/workflow.md) for exact route, prompt, confirmation, polling, and recovery details. ## Decisions that require confirmation Planning, post writing, and prompt drafting are free. The optional Xiaohongshu lookup is the one thing that can charge before generation, and it is priced and approved on its own. Before any image generation or revision, obtain one clear confirmation of the frozen lookbook card, every paid image request, source and reference order, canvas, model, controls, count, current price, maximum charge, and total call count. Every changed outfit, source or reference order, prompt, slide role, canvas, model, count, output relationship, or control is a new paid request with a new confirmation and a new `client_request_id`. A focused revision to an accepted slide is also fresh paid work. ## Execute and deliver Use only this package's bundled `scripts/mcp_client.py` for every remote operation. Send one JSON object on standard input after `call <tool-name>`. Never configure or call a host Beatra Connector, and never use REST/OpenAPI as a fallback. Read [Bundled MCP Client diagnostics](references/mcp-connection.md) for commands and connection troubleshooting. For source-photo routes, upload the local file and put its returned artifact at `images[0]` in the transform request. Later images may guide composition, palette, setting, or styling in the stated order. For a concept-only route, use `beatra.images.generate` with the selected vertical canvas. For a focused revision, use the accepted slide as `images[0]` with `beatra.images.edit`. Register the package through `beatra.installations.register` on first use. A returned `task_id` belongs to the original approved work: poll it only with `beatra.tasks.get`. When a create response is genuinely unknown, retain the same frozen payload and `client_request_id`; if the task ID is missing, use `beatra.tasks.list` and verify a candidate with `beatra.tasks.get` before any replay. Use `beatra.tasks.cancel` only at the user's request. If cancellation returns `409`, keep polling the original task and report cancellation only after its terminal `status` is `canceled`. Review accessible images against the approved outfit must-keeps, slide role, 3:4 composition, visual continuity, and title-safe placement. Deliver only facts returned by the completed tasks: artifact links, dimensions, format, resolved model, task IDs, and `billing.net_charged_credits`. Present the ordered lookbook, cover and slide roles, title ideas, caption beats, tag set, and any visible drift that matters to the user's next revision. ## Reading Xiaohongshu before you write Optional, and paid. When the connection exposes Beatra's public social lookup, this Skill can read Xiaohongshu directly instead of working from what the user remembers: one page of notes matching a keyword, one specific note the user pastes, one page of that note's top comments, and an account's profile or recent notes. Six operations, Xiaohongshu only. **Every one of them costs 60 credits**, and there is no cheap operation on this platform to fall back on. The same reads cost 6 on TikTok. They cost 6 on Douyin too — except Douyin's own keyword search, which is also 60, so do not say "ten times Douyin" without naming the read. A three-step read — the field, one note, that note's top comments — is 180 credits, and every further page is another 60. Say the number before offering anything, confirm each lookup on its own before it runs, and say plainly that this Skill's own deliverable arrives either way at no cost. Offer one read, not a plan of four. The rule is the whitelist, not a list of exceptions: a platform with no operation on it cannot be looked up from here, and another platform's notes are never presented as Xiaohongshu's. A returned image URL is not a viewed image — state a visual finding only about an image the host can actually open. Every figure that reaches the work is labelled as looked up with the time it was read, or as supplied by the user, or as missing. Nothing is estimated, and nothing is carried in from what notes in this category usually do. See [reading Xiaohongshu](references/note-lookup.md) for the operations, the argument routes, the confirmation wording, and how a result is reported and recovered. ## References by task - For the lookbook card, four-slide defaults, source-reference roles, and post angle: [lookbook planning](references/lookbook-planning.md). - For precise transform, concept, edit, confirmation, task tracking, recovery, and review: [lookbook workflow](references/workflow.md). - For first use or expired credentials: [installation and authentication](references/installation-and-auth.md). - For the non-billable installation registration: [installation registration](references/installation-registration.md). - For task fields and polling: [tasks and results](references/tasks-and-results.md). - For balance, validation, and structured errors: [billing, errors, and recovery](references/billing-errors-and-recovery.md). - For client commands and diagnostics: [Bundled MCP Client diagnostics](references/mcp-connection.md). - For removal: [uninstall and disconnect](references/uninstall-and-disconnect.md). ## Updates and security This Skill runs only the files installed with it and leaves them unchanged while it runs; it never downloads or executes other code. It talks to Beatra's official services over HTTPS, and a file you choose to upload goes only to the one-time upload address Beatra issues for it. Once a day at most, the bundled client reads Beatra's public version list. When the store or tool you installed from has a newer version, it prints one line with the command that updates it there. Updating stays your choice, through that store or tool. See [updates and security](references/updates-and-security.md).

存档时间线

版本索引永久保留；内容副本只保留最新 1 份，更早版本仅留索引与哈希（存档时间线的证据链不会因此断裂）。

下载存档内容副本