Uses official DeepSeek APIs so an agent can search, read links or files, run OCR, transcribe or synthesize speech, and answer with Flash, then return checkable results.

向阳乔木
@vista8
喜欢摇滚乐、爱钓鱼的PM 第二届GEO大会报名-11-21/22日-深圳南山 https://t.co/N75cblhK9R
게시물
10Scaffolds a one-shot native macOS app with Lucide icons, settings/About, Sparkle updates, notarized DMG packaging, and GitHub Releases.
Opinionated frontend design advisor that fights generic AI UI: style fitting room plus dozens of real-site design systems for Claude Code.
Turns event briefs, URLs and brand assets into commercial registration posters with verified QR codes and Chinese/Latin typography checks.
qiaomu-ds-search wraps DeepSeek's Search and Fetch endpoints for sourced web research and page reading, including WeChat articles, at 1,000 calls for 5 yuan.
qiaomu-screen2video turns an existing screen recording into a finished video with bilingual subtitles, highlighted keywords, meme stickers and light music, plus horizontal and vertical covers and an end card.
anatomy turns a one-line machine description into an interactive isometric figure with rotating parts and WebGL shaders; the post suggests borrowing its acceptance, data and visual-element rules for your own Skill.
Use qiaomu-codex-imagegen to empower any agent with Codex's built-in image generation for high-quality covers, posters, and infographics.
Use qiaomu-obsidian-dev to develop, debug, and release Obsidian plugins with native APIs and proven engineering practices.
Generate professional product promotional videos from a GitHub repository link and a one-line description using the qiaomu-cut skill.