Agent Skills
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
Detects flame and smoke in videos and images, returning structured reports, anomaly markers, and cloud history queries for security, forest, and industrial alerting.
An eight-dimension life planner for assessing status, setting goals, analyzing gaps, and outputting an action blueprint.
A short-video batch production workflow for Douyin, Kuaishou, Video Account, and Xiaohongshu, covering ideation, scripting, asset creation, editing, distribution, and analytics.
Generate self-contained, warm-toned single-page HTML reports with inline SVG charts for weekly, progress, and postmortem summaries.
Call the Lux3D China API for image-to-3D, text-to-3D, material redraw, four-view, export, and task queries.
Provides A-share financial red-flag audit task lists, dependency topology, and metric requirements for Universal Task OS execution.
Assists patent invalidation requests by decomposing claims, searching prior art, organizing evidence chains, and drafting request documents and evidence lists.
An AI-agent skill for local browser control via the 127.0.0.1:10086 daemon, covering navigation, snapshots, clicks, fills, and screenshots with user/task tab separation.
Convert a single PDF, Word, or Excel document into an offline single-file HTML with synchronized text and charts, style presets, editable data re-rendering, and export options.
For spoken short-video creation, it generates step-limited scripts, evaluates existing scripts, and plans content matrices.
A design-engineering capability for polished visual front-end artifacts, with fact-checking, design context, and refined HTML pages, prototypes, decks, and visualizations.
Guides interactive knowledge exploration of visual novels through works, production, and fandom from an otaku researcher perspective.