bibidu avatar

bibidu

Community

@bibidu · beijing

54Followers
|
21Public Repos
|
6Published Skills

Building tools for developers. Mini-app infrastructure · Cross-platform · AI agents. Based in Beijing.

Skills Distribution
DomainGrowth, Mark...Short-Video Produc.. (40%)Social Media Accou.. (25%)Cloud Object Stora.. (20%)Static Site Deploy.. (15%)

Agent Skills by bibidu

Showing 6 vetted skills indexed across 1 GitHub repositories.

Frequently Asked Questions About bibidu

FAQPage Schema
What tasks can I accomplish with bibidu's skills?

You can produce 9:16 Douyin fengshui videos with cloned voiceovers and subtitle gates, run two Douyin account manager personas (fengdashi, shedashi) for topic selection and scheduling, upload images to Aliyun OSS with signed URLs, and deploy an Era frontend to a self-hosted Aliyun server with Caddy reload.

Who are bibidu's skills designed for?

They target Chinese social media operators and developers running Douyin content accounts, particularly fengshui and AI-productivity channels, plus engineers maintaining a self-hosted Era site on Aliyun lightweight servers who need repeatable deployment and OSS media upload procedures.

How does the fengshui vertical video production flow work?

Provide a Douyin or video link; the skill extracts the Chinese voiceover, rewrites under 5%, selects elder voice cloning via CosyVoice, VoxCPM2, or VoxCPM 0.5, enforces word-level and audio-quality gates, then requires 10-second preview and first-frame confirmation before assembling the 9:16 final cut.

What prerequisites and dependencies do these skills require?

Video skills need CosyVoice/VoxCPM voice cloning models and Douyin source links. OSS upload requires an Aliyun OSS bucket and oss-upload.sh. swas-deploy requires SSH access to server 39.106.179.17, a git main branch, npm, and Caddy. Account managers require Feishu notification access.

Are any bibidu skills deprecated or restricted?

Yes. The tuwen multi-page image-text skill is fully deprecated; fengshui and yangzhai content must route to the fengshui vertical video skill instead. Era stacked-text covers and gc-minimal multi-page image output are explicitly forbidden across all current skills.