laohan-douyinsousuo

A unified content creation engine for Chinese video creators that transforms multiple input types (screen recordings, URL queues, trending topics, outlines, raw text) into oral broadcast scripts with cover image prompts. Features audio extraction, speech-to-text transcription, trending topic analysis, and customizable writing styles.

A Douyin (Chinese TikTok) content compliance checking skill that scans scripts for seven categories of violation risks including traffic diversion, extreme claims, medical/financial promises, and platform-specific restrictions, outputting structured reports with replacement suggestions.

An AI trending topics aggregator skill that fetches from three parallel sources (AIHOT curated feed, opencli 9-platform hot lists, and Douyin AI filter), then merges, deduplicates, and cross-validates results into a comprehensive daily briefing. Designed for Chinese-language AI news coverage and content topic discovery.

A unified content calibration skill that automatically routes user intents to appropriate actions including scoring, prediction, review, and status tracking for content quality assessment in Chinese content creation workflows.

An all-in-one internet content acquisition skill that handles video downloading, audio extraction with transcription, subtitle downloading, comment collection, and web scraping across 30+ platforms including Douyin, TikTok, YouTube, Bilibili, and Xiaohongshu. Uses a degradation chain approach, automatically selecting the optimal tool from lightweight HTTP requests to browser automation.

A skill for generating professional storyboard image prompts optimized for diffusion models like FLUX, SDXL, and Gemini, with quality validation and automatic splitting into individual prompt files for product-focused short video production.

Generates Gemini image prompts for video thumbnails in the Qiuzhi 2046 style (white background, symmetrical bold text, exaggerated props), dynamically creating 3 visually distinct style variants based on script content. Outputs 9 total prompts across 3 aspect ratios for Douyin, Xiaohongshu, and Bilibili.

A screen recording automation skill that converts narration scripts into bash recording scripts, orchestrating CLI operations via tmux and browser operations via Playwright, then recording the physical screen with ffmpeg to output 1080p MP4 video.

A specialized skill for creating, modifying, and optimizing Claude Code agent skills, incorporating best practices from anthropics/skills, mattpocock/skills, and 9arm-skills with a focus on precise trigger descriptions, phase gating, and minimal yet effective skill structure.

Converts voiceover scripts into slide images using NotebookLM to generate Chinese-language slides, then exports them as PDF-to-PNG for direct use in CapCut (Jianying). The workflow handles notebook creation, source upload, slide generation, and image extraction automatically.