Rising stars
Entropy-TOPSIS-ML + Reviewer Objection
Produces a complete Methods + Results report from your entropy weighting + TOPSIS ranking + ML classifier (XGBoost, Random Forest, Logistic Regression, etc.) outputs. Correctly classifies AUC interpretation (Hosmer & Lemeshow criteria) and, if the train/test performance gap is large, states the overfitting risk explicitly. It also pre-identifies at least 4 objections journal reviewers frequently raise (e.g., "how was class imbalance handled," "no external validation set was used") and suggests a defense sentence for each. Suitable for: researchers working in health informatics, clinical risk prediction, or hybrid MCDM-ML methodology who want methodological robustness before journal submission.
TOPSIS Decision Analysis Writer
Produces a complete manuscript-ready Methods + Results text from your TOPSIS analysis outputs (normalized decision matrix, distances to positive/negative ideal solutions, closeness coefficients). Provides a ranking table and interprets which criteria the best and worst alternatives excelled or lagged in. Suitable for: researchers working on multi-criteria ranking problems such as supplier selection, product/service comparison, or clinical decision support.
AHP Criteria Weighting Writer
Converts the user's AHP (Analytic Hierarchy Process) pairwise comparison matrix outputs (criteria weights, consistency ratio) into an academic "Criteria Weighting" section ready to drop straight into a manuscript. If the consistency ratio (CR) exceeds 0.10, it states this openly rather than hiding it. Suitable for: researchers who use MCDM methods such as TOPSIS/VIKOR and determine criteria weights with AHP, and graduate students.
Real People × Line Art
Keep the real people from the photo, drawing restrained colors from their appearance while turning the surrounding environment into playful line art with bold strokes, minimal detail, and a from-memory feel.
XHS Note Growth Studio
Turn a topic, product materials, real experience, or scattered ideas into a complete XHS content plan: first clarify the audience, core promise, and differentiated angle, then produce three title options, a clearly structured post, relevant tags, a 3:4 cover visual brief, and a pre-publish quality check. It also supports competitor inspiration, rewrites, title optimization, and cover-only mode. When in-platform data isn't available, it will say so clearly rather than fabricate engagement numbers or trends; it won't copy others' content or publish on the user's behalf without confirmation.
AI Storyboards: Short Dramas
Enter the script, novel, or client-provided script here, but be sure to simplify it first. Many scripts include editing instructions or opening and closing sequences that belong to post-production; remove those elements. This tool is for pre-production. If seedance2 is also asked to handle post-production, the output quality may decline.
Product Blueprint | Dual Maps
Turn product ideas, PRDs, meeting notes, or documentation of an existing system into cross-validated product information architecture, core user flows, system context, and data flow diagrams. It also produces a product-to-system capability map, key risks, and open decisions. Ideal for product managers, founders, solution architects, and engineering teams conducting requirements reviews, planning SaaS/AI products, designing complex back-office systems, and modernizing existing systems. Supports Chinese and English input, prioritizes generating shareable responsive web pages, and preserves editable Mermaid source code.
AI Video & Storyboard Prompts
Turn storylines, character profiles, or existing video prompts into professional storyboards and ready-to-use shot prompts for AI video generation. Whether the scene involves contemporary life, historical drama, mystery, romance, or high-payoff drama, the skill focuses on character relationships, pacing, and emotional hooks. It clearly defines shot scale, camera position, perspective, camera movement, composition, lighting, ambient sound, and dialogue, giving every shot a clear visual focus and actionable movement. The skill pays close attention to character consistency and scene continuity. It helps lock in each character’s appearance, clothing logic, identity and status, and defining details, reducing generation issues such as facial changes, visual drift, spatial confusion, and unclear actions. A character’s gaze, expressions, breathing, hand movements, and physical reactions are also translated into visible performance changes, strengthening conflict, tension, twists, and emotional progression. You can start with a complete storyline or provide existing character assets, storyboards, or prompts for targeted optimization. When you already have an established character profile, it can be reused directly to avoid redesigning it. The complete sequence can then be organized into clips of 15 seconds or less for generation with mainstream AI video models, while preserving visual, audio, and duration details for easy copying, use, and further editing.
Sow Herd Index Overview
Turn a sow herd breeding-index Excel file into a clear overview of herd-level breeding performance, helping pig farms quickly compare the overall results of the nucleus herd, multiplier herd, and entire herd. The Skill groups records by herd type, summarizes the sample size and average 120 kg maternal-line index, shows high- and low-score tier performance, and calculates the average reproductive index and 120 kg paternal-line index. It is suitable for routine breeding analysis, herd evaluation, and periodic reporting. Upload a breeding-index Excel file for the entire herd to receive a summary and generate an Excel overview that can be reviewed and used further. Statistical definitions and data columns can be adjusted to fit different farm data structures. If the file includes herd categories beyond the nucleus and multiplier herds, the Skill will prompt you to confirm whether they should be included. In addition to the overall overview, the results include detailed statistics for the maternal-line index, making it easier to understand differences in breeding levels within the herd.
Swine Breeding Index Tool
Automatically analyze and export swine breeding indexes for breeders who need to process sow herd data in the GBS breeding software. You can calculate the age at 120 kg, 120 kg sire-line index, 120 kg dam-line index, and reproductive index—or generate multiple breeding indexes and estimated breeding value (EBV) results for the herd at once. The Skill prepares data and performs BLUP calculations based on your farm, breed, testing range, birth dates, analysis model, and selection scenario, then filters the results according to specified conditions. It supports common scenarios such as dynamic core herd balancing and replacement pig selection, helping you identify breeding animal records that meet business requirements. After confirming that the number of filtered results is reasonable, you can export the analysis table to avoid sending insufficiently screened data into subsequent workflows. The final output is a clearly named Excel file containing breeding analysis fields such as herd rank, breeding animal ID, reproductive index, 120 kg sire-line index, 120 kg dam-line index, and related EBVs. It also reports the save location, record count, and selection scenario used. This Skill is suitable for sow herd breeding evaluation, replacement pig selection, core herd adjustment, and routine breeding data organization.
ImageXiaoxiaodong x P086
Panel 086 keeps only the 3–5 visual cues that make an image work. It reshapes the subject with silhouettes, geometric forms, and simple color blocks, then uses 2–4 spot colors drawn from the photo, along with dry-brush, brayer texture, dragged ink, broken ink, and uneven coverage to create a medieval modernist screenprint. The composition always seeks one clear focal point, one primary direction, and one meaningful area of negative space. Best for When you want more visual order in a photo: Keep only 3–5 key visual cues, using silhouettes, geometric forms, and color blocks to establish a clear focal point. When you want an authentic printmaking texture: Let dry-brush, brayer texture, dragged ink, broken ink, and uneven coverage become part of the visual language. When you want vivid but restrained color: Compress the most dynamic relationships in the original image into 2–4 spot colors, allowing the paper base and negative space to actively shape the composition. When you need flexible output: Supports top-to-bottom, left-to-right, pure design, multiple aspect ratios, and four-edge wallpapers, as well as batch processing for image directories.
ImageXiaoxiaodong x Line x Dot P084
Panel 084 uses fine black or dark lines to distill the subject, buildings, roads, and spatial relationships. Vertical, horizontal, angled, curved, and perspective guide lines establish the geometric structure. Stippling／dotted texture then uses shifts in density to shape the sky, ground, distant scenery, tree shadows, lighting, and the atmospheric layers around the subject. Complex backgrounds are intentionally reduced, leaving a clear focal point, generous negative space, and a few accent colors with clear sources. Ideal for these situations When cities, architecture, and spatial relationships need to read clearly: Use geometric linework, perspective guide lines, and key contours to refine roads, buildings, and scene structures. When you want to show air, light, and depth without covering the image in texture: Use gradients in stippling／dotted texture density to create quiet layers. When you prefer a restrained color palette: Use blue-gray, off-white, and warm white as the main colors, with only a few spot color accents from the original image to highlight the focal point. When you need flexible delivery: Supports top/bottom, left/right, design-only, multiple aspect ratios, and four-sided wallpaper, and can also process image directories in batches.
ImageXiaoxiaodong x Doodle x P083
Panel 083 first reevaluates the photo, keeping only one subject worth remembering and at most one or two necessary environmental cues. It then deliberately distorts a proportion, action, expression, or structural element, making it crooked, clumsy, and slightly absurd. Trembling outlines, intentionally inaccurate proportions, rough crayon textures, and generous white space create an Ugly-cute naïve doodle; the subject’s identity must remain accurate, and the humor must come from the original image rather than arbitrary added jokes. Good for these situations A photo contains a memorable personality moment: Find humor in an action, expression, proportion, or contrast instead of mechanically tracing the image. You want ugly-cute without cheap sweetness: Draw things deliberately crooked, clumsy, and slightly absurd while keeping the subject immediately recognizable. You want a complex scene to become simple and precise: Keep only one subject and at most one or two environmental cues, using rough coloring and generous white space to strengthen the memorable detail. You need flexible delivery: Supports top-to-bottom, left-to-right, design-only, multiple aspect ratios, and four-device wallpapers, as well as batch processing for image directories.
Game UA Creative Teardown
Analyze game user acquisition creatives, including Guangdiantong/Tencent Ads, from two perspectives: the seven-module emotional psychology framework and the four-part campaign diagnosis framework. First, complete a factual creative checklist and grade the evidence. Then analyze emotional psychology—first-three-second hook, emotional curve, closing call to action, monetization points, psychological needs, human insights, and iteration direction—before applying C-E-B-A labels and a seven-layer diagnosis to identify problems and turn them into campaign actions (targeting, bidding, and a promise-reduction SOP). The final output is a nine-section teardown report. Core principles: conclusions at the creative level must be supported by A-grade visual-frame evidence; dual-layer conclusions must be based on actual user experience; users brought in by the creative must not be treated as the game's true target audience; and the main text should use plain campaign-buying language.
Weaning Sow Core Group Review
Dynamic core-group balance analysis after weaning. Based on weaned sow ear tags (images, handwriting, or voice) and the provided breeding index data for the full herd, this Skill verifies ear tags, checks statuses, calculates metrics (maternal-line index, reproductive index, paternal-line index, and parity), determines whether sows should be culled or downgraded, and generates an Excel dynamic core-group balance sheet. Use this Skill when the user has provided weaned sow ear tags and a breeding index data file, such as 20260827母猪群育种指数.xlsx, and mentions “adjust the core group,” “adjust core-group sows after weaning,” “dynamic core-group balance,” or “weaned sow adjustment.” If the user provides only ear tags without the data file, use the weaning-core-balance-auto Skill for end-to-end processing.
Spark | Video Diagram Maker
Many videos don’t lack content—they’re simply too abstract or dense. Viewers understand one sentence, then forget it by the next. “Spark | Video Diagram Maker” turns your ideas, brainstorms, outlines, scripts, transcripts, or articles into visual concepts worth showing. It then determines whether each idea is best expressed as a process, comparison, architecture, hierarchy, loop, branch, decision tree, or another diagram structure. It won’t draw just to meet a quota, or cram complex content into one cluttered infographic. Multiple diagrams maintain the same visual language, naming, and semantic color system, making them easy to insert into a video as a visual framework for explanation. Especially useful for: Explaining complex concepts Breaking down workflows Comparing two methods Showing system structures Clarifying cause-and-effect relationships Turning long brainstorms into a set of visual nodes for a video The default output is a 16:9 static diagram. If the runtime environment cannot provide an editable Excalidraw source file, this will be stated clearly rather than presenting a static image as a .excalidraw file. Use cases Use case 1 | Turn brainstorms into diagrams You have a long stream of spoken ideas but no complete script. The skill identifies the points most worth illustrating and organizes them into a set of diagrams with a clear narrative sequence. Use case 2 | Turn scripts into video visual nodes You already have a complete script but don’t know where to add visuals. The skill determines which sections are worth visualizing and selects suitable diagram types. Use case 3 | Explain complex concepts For example: how AI Agents work, why workflows matter more than one-off demonstrations, how a business model operates, or which layers make up a product architecture. Use case 4 | Break down workflows and processes For example: topic pool → research → Hook → script → production → review. Use case 5 | Make comparisons For example: without a workflow vs. with a workflow; traditional methods vs. AI methods; one-time success vs. stable, repeatable results.
Spark | Video Hook Builder
Many videos aren’t held back by weak content—their openings simply fail to make the real value clear. “Spark | Video Hook Builder” follows the BRENS Hook Builder approach: first identify the Content Promise, then determine which of Big, Relatable, Easy, New, and Safe are genuinely present before writing the Hook. It won’t invent data, experiences, results, or authority just to make the opening feel “more explosive,” and it won’t force all five BRENS dimensions into every Hook. You can use it to refine an existing opening or build one from scratch using a script, transcript, or outline. If you’ve already analyzed a reference video with “Spark 02,” you can also transfer only its structure to your own content. Core sequence: Content Promise → BRENS → Hook Ideal for creators making YouTube videos, video podcasts, educational content, hands-on reviews, tutorials, interviews, and other formats where more efficient opening communication matters. Mode A｜You have an opening, but it isn’t engaging enough (Polish) Provide: your current Hook + the main content, outline, or content summary. Example: Here’s my current opening and the video content. Keep my tone as much as possible, make only necessary edits, and bring the genuinely worthwhile parts forward. Mode B｜You have the content, but no opening yet (Build) Provide: a transcript, complete script, outline, or clear content summary. Example: Here’s my video content. First determine what this video can genuinely promise, then write 1 recommended opening using BRENS. Mode C｜You’ve already analyzed a reference video with Spark 02 (Transfer) Provide: Spark 02’s Hook Outline or reusable template + your own content. Example: Transfer only the beat functions from the reference structure below. Don’t copy the original video’s facts or wording; use my content to write my own opening. Mode D｜You only want a diagnosis, not an AI rewrite (Audit) Provide: your existing Hook + the main content or outline. Example: Only diagnose it with BRENS—don’t rewrite it. Tell me the 1–3 changes that matter most right now. Recommended input priority: Finished-video transcript / complete script Content outline Clear content summary Existing Hook / title Reference structure from Spark 02 The closer your materials are to the finished video, the less likely the Hook is to “outpackage” the content. If your target audience, platform, video length, tone, number of versions, or desired degree of change is already decided, include it directly.
ResearchSpark | Video Hook Analyzer
When preparing to make a video, we often run into the same problem: you can clearly see that a reference video “knows how to tell a story,” but when it’s your turn to create, you can only imitate it by feel. 「Spark | Video Hook Analyzer」is designed for this step. Give it a specific YouTube video. It first verifies the video and reads captions that can be checked, then separates three things that are easy to conflate: Literal 30s: What actually happens from 0:00–0:30. Intro: The entire opening before the main content begins. Hook / Re-hook: One or more structural beats that genuinely work to keep viewers watching. So it won’t force the first 30 seconds of every video to be interpreted as the Hook just because “everyone says the first 30 seconds matter most.” If the real Hook comes after the one-minute mark, it will tell you plainly. If the captions or timestamps are insufficient, it will clearly identify the evidence gap instead of filling it with an answer that merely looks professional. After breaking down the opening, it continues by using an outline to reconstruct the structure of the full video: Does it move forward through a timeline, experiment, process, comparison, story, chain of questions, or another organizing device? Finally, it abstracts these structures into [slots] templates. The templates preserve the sequence and function of the structure—not the original creator’s experiences, wording, brand, numbers, opinions, or authority. What you ultimately get is not a video summary, but a Hook Outline + full-video structure template that can be handed off to the next stage of creation. What it does not do It does not: Write the final new Hook for you; Score or rewrite your opening with BRENS; Search the entire niche or channel at scale; Generate a complete script; Generate diagrams, assets, animations, or thumbnails. Finding reference samples at scale belongs to the upstream 「Spark | YouTube Content Researcher」; refining your own Hook belongs to the corresponding follow-up Skill.
Architectural Woodcraft
Transform any uploaded architectural reference image into an ultra-realistic photograph of a handcrafted natural-wood architectural scale model on a model-maker's workbench, preserving the original design.
STORYBOARD DIRECTOR GPT
The key principle of this scene is not to use multiple camera angles merely for aesthetic purposes. Every change of angle must serve a function: to clarify the space, increase tension, accelerate the scene, show a reaction, or confirm the outcome of an action.