Agent Skills · Design & Media
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
Use Composio MCP to connect HeyGen, browse templates and avatars, generate template-based videos, poll status, and retrieve shareable links.
Calls MiaoYan XiaoZhi PicTech.cc background removal to turn local files, URLs, or folders into transparent PNGs or white/black background product images for cross-border listings.
An intelligent background-removal and enhancement tool for portraits, products, and game characters, with local/API matting, hair refinement, multi-subject extraction, upscaling, sharpening, and batch transparent PNG output.
An LLM context compression wrapper that reduces token usage by 60%–95% and includes Python/TypeScript SDKs.
Protected brand visual identity design capability that must pass WorkBuddy authorization via `auth.py` and retrieve cloud protected instructions before running.
Extract spoken scripts from local MP4 files or video links into TXT without burned-in subtitles; requires Whaleclip desktop login and VIP.
Use local ComfyUI with the Z-Image Turbo model to generate 1024×1024 images from text prompts, with service, model, and dependency checks.
Wraps common yt-dlp commands for downloading videos, extracting audio, managing playlists, subtitles, proxy settings, and troubleshooting.
Creates and polls Baiyin open-platform image generation tasks with fixed models, resolution, aspect ratio, and reference image upload support.
Turns finalized copy, scripts, or report points into mobile-readable social cards with 77 brand styles, HTML layout, and optional AI assets, then renders publishable PNGs.
Generate health-science posters via Volcano Engine Seedream, with 120+ built-in topics, configurable organization name, knowledge base, and output path, plus image, JSON, and text outputs.
Provides Anthropic brand colors, typography rules, accent guidance, and fallback fonts for consistent visual formatting.