skills

Use when generating believable editorial young-adult human portraits or fashion/UGC character images with Nim, especially Gen-Z fashion casting, model/person look development, aesthetic human references, or prompt-driven human generation. Asks for age or age range, sex/gender presentation, ethnicity/look, and additional notes; defaults to Nano Banana Pro in 9:16, but supports user-specified models such as GPT Image 2 by discovering the live Nim model contract.

Use when improving, writing, rewriting, planning, or executing AI video prompts for Seedance, Higgsfield, Nim MCP, UGC, cinematic generation, image-to-video, reference frames, character consistency, GPT Image 2, Nano Banana Pro, 2K stills, LUTs, camera quality, shot lists, storyboards, beats, montage, sequence, camera angles, or visual grammar.

Use when writing, improving, planning, or generating Nim/Seedance UGC video prompts for TikTok, Reels, Shorts, creator ads, direct-to-camera hooks, selfie videos, iPhone footage, testimonials, street interviews, mirror try-ons, outfit changes, product/location walkthroughs, VO, lip sync, audio references, or image-to-video prompts with character/location/product references.

Use when the user wants commercial product visuals or videos from product images, reference assets, or a product brief: ecommerce listing images, product photoshoots, studio/catalog shots, white-background product photos, lifestyle scenes, ad creatives, hero images, product launch visuals, social content, UGC-style creatives, on-model apparel/accessory shots, Shopify/Amazon PDP assets, product-in-use scenes, product videos, or similar selling, merchandising, marketplace, brand, or campaign imagery. Search Nim templates first for product/photoshoot/ad-style workflows, show the user matching template options with cost/input info, get an explicit template choice or approval, inspect the selected template contract, collect required inputs, upload references, and run the template. Fall back to Nim model generation only when no suitable template exists or the user explicitly asks for raw generation.

Improve image or video quality with Nim upscalers. Use when the user asks to make an existing image or video better, higher quality, sharper, clearer, cleaner, higher resolution, less blurry, less pixelated, enhanced, restored, upgraded, upscaled, 4K, HD, 60fps, smoother, or otherwise improved in visual quality without requesting a creative restyle or new generated content. Routes images to upscale_image and videos to upscale_video. Uses bytedance by default for video upscaling and asks for missing source/target details before running.

Create new Nim images of an existing character, person, product, object, mascot, avatar, or branded item while preserving visual identity across scenes, poses, lighting, outfits, or artistic media. Use for character consistency, consistent character images, identity-preserving edits, product consistency, same person/object in a new scene, or variants from one or more reference images. Discovers the current Nim model with models_explore, follows its generationContract exactly, uploads references with media_upload, and submits with generate_image.

Use when the user asks for a video longer than Nim's single-clip cap — a duration roughly between 15 seconds and 5 minutes ("1 minute video", "30 second video", "90 sec clip", "2 minutes long"), or asks for a format that implies more than one clip's worth of content without giving an exact number: "YouTube Shorts", "Reels", "TikTok", "a short film", "an episode", "a series". Nim's Seedance video generation is maximum 15 seconds per clip. This skill is NOT for single clips of 15 seconds or less — use nim-generate / nim-b-roll-seedance / nim-hookgen-seedance for those.

Use when creating product or lifestyle b-roll videos from a user brief with optional product, character, and location reference images, especially when the output should be generated with Nim Seedance. Handles brief intake, missing-input checks, reference mapping, Seedance prompt drafting, approval checkpoints, Nim upload/generation, and iteration.

Check Nim credit balance, explain insufficient-credit generation failures, and help the user buy or upgrade Nim credits through the Nim MCP. Use when the user asks about credits, balance, pricing packs, buying more credits, upgrading, or when generate_image, generate_video, or run_template returns status="insufficient_credits".

Generate or edit images and videos with Nim. Use when the user says "generate an image", "make a picture / poster / logo / wallpaper", "edit / restyle / remix this image", "make a video", "animate this photo", "image-to-video", "turn this into a clip", "give me variations / options", or asks for any visual asset from a prompt or a reference. Discovers the right Nim model with models_explore, follows its generationContract exactly, uploads references with media_upload, and submits with generate_image / generate_video. Supports text-to-image, image editing, text-to-video, and image-to-video. NOT for: code, text/chat answers, or audio-only tasks.

Author a production-ready Seedance 2 hook video for short-form UGC (TikTok / Reels / Shorts) and generate it through the Nim MCP. Use when the user wants a talking-head "hook" ad/UGC clip from a product brief, asks to "make a hook video", "generate a UGC hook", "spoken hook ad", or "Seedance hook". The agent acts as the prompt-writing LLM: it intakes the brief, proposes 5 spoken hook options, collects character/product/location reference images, assembles the Seedance brief, rewrites it into a final Seedance prompt, and generates the video on Nim Seedance 2. NOT for: long scripts, code, text-only answers, or audio-only tasks.

Find, inspect, and run Nim templates through the Nim MCP. Use only when the user explicitly asks for Nim templates, says "template" / "шаблон", asks to browse or recommend templates, or asks to run a specific named Nim template. Uses explore_templates, get_template, media_upload, and run_template. NOT for ordinary image/video generation requests; use nim-generate for those.