AWorld

Use for framework-gated self-evolve workflows in AWorld: evolve skills, create trajectory-backed proposals, inspect self-evolve run artifacts, run aworld-cli optimize, or prepare verified apply decisions through aworld.self_evolve gates.

Draft skill generated by self-evolve for trajectory-backed task handling.

An intelligent assistant specialized in handling media files (images/audio/video). **Only for media file analysis**, does not handle document types. Media files that can be processed: - Images: .jpg, .jpeg, .png, .gif, .bmp, .webp, .svg - Audio: .mp3, .wav, .m4a, .flac, .aac, .ogg - Video: .mp4, .avi, .mov, .mkv, .webm, .flv Files that cannot be processed (please do not trigger this skill): - Documents: .pdf, .doc, .docx, .txt, .md, .rtf - Spreadsheets: .xlsx, .xls, .csv, .tsv - Presentations: .pptx, .ppt, .key - Code: .py, .js, .ts, .java, .cpp, .go, .rs - Archives: .zip, .tar, .gz, .rar, .7z - Executables: .exe, .bin, .app, .dmg - Databases: .db, .sqlite, .sql - Configuration files: .json, .xml, .yaml, .yml, .toml, .ini - Web pages: .html, .htm, .css Trigger conditions: When the user explicitly requests to analyze image/audio/video content, or when the file extension belongs to the aforementioned media types.

X (Twitter) 抓取 skill - 通过 agent-browser (CDP) 抓取指定用户推文或首页推荐流，支持关键词过滤、Tab 切换、多格式输出。使用场景：按用户/关键词抓取时间线、查看首页推荐流、生成 RSS/JSON/Markdown。

Legacy demo skill for plugin discovery tests.

Create ad-ready product video from product images, with or without character/subject images. The workflow leverages AI-powered image composition, scene understanding, and video generation. Video prompts should follow commercial shot language—visual hooks, product presence, hero shots, detail showcase, function expression, and dynamic visuals.

Create ad-ready product images (single or collage) by back-solving sub-image sizes from target output ratio, grounding scene design with media_comprehension, generating images via image_generator with strict request params and actor-count control, and pairing each deliverable with a short social tagline for 小红书/抖音.

Single-file TikTok/Douyin video download and traffic metrics via TikHub API using only httpx; optional persisted raw API JSON plus a stdlib post-processor emitting CSV and simplified JSON. Supports one URL or concurrent batch (max 10 workers). No dependency on any project codebase.

Lightweight TikHub Xiaohongshu image-search workflow. Prioritizes single-request usage with curl or minimal Python, saves raw API JSON by default, and includes a small stdlib post-processor for CSV and simplified JSON. Use when the user wants Xiaohongshu keyword image search, page-based pagination, or structured note/image metadata from TikHub without a heavy wrapper.

Lightweight TikHub YouTube search and video-detail workflow. Prioritizes single-request usage with curl or minimal Python, saves raw API JSON by default, and includes a small stdlib post-processor for CSV and simplified JSON. Use when the user wants YouTube comprehensive search results, continuation-token pagination, or structured video metadata from TikHub without a heavy wrapper.

Search and summarize the latest 7 days of AI news and X discussions using public sources plus browser-based X collection. Use for recent AI news, trends, X discussions, industry briefs, and summaries organized into hot topics, viewpoints, and opportunity areas.

Prevent stutter and frozen frames when embedding a child video inside a parent in code-driven pipelines (Remotion, After Effects scripting, FFmpeg filter graphs). Explains why sparse keyframes break frame-accurate seek during per-frame export, and how re-encoding with H.264 all-intra GOP (-g 1) and yuv420p makes every frame independently decodable. Includes FFmpeg command, parameter notes, file-size tradeoffs, and a reusable rule for any seek-heavy programmatic video workflow.

Core storytelling rules for AI video scripts: concrete metaphors instead of abstract jargon, the mute test (story reads without audio), visual contrast and closure, physically visible causes of failure or success, visualizing the “eureka” beat, camera motion tied to physics, in-scene transitions instead of black cuts, character consistency and multi-speaker action/lip sync timelines, and three-act pacing with Mandarin VO speed (~4.5 chars/s) and breathing room for action and SFX.

Standard operating procedure for automated AI video production using a Remotion (code) and diffusion (model) hybrid pipeline. Covers narrative DNA (hero, show-don’t-tell, three-act arc), technical specs (duration, integer segment lengths, resolution, fps, Mandarin pacing), tech-selection matrix (diffusion vs code), a five-part diffusion prompt protocol (style, micro-timing, entities, camera, transitions), end-to-end execution workflow, and a fixed output template (metadata table + per-shot table). Complements create-video and Remotion best-practice skills for execution quality.

Burn hard subtitles from UTF-8 SRT files using moviepy 2.x with CJK-capable system fonts; tune font size, placement, stroke, and encode settings (bitrate or CRF) to avoid oversized outputs. Documents ffprobe/ffmpeg workflows for inspection, encoding, and batch jobs; troubleshooting for fonts, bitrate, and pacing. Covers voiceover with edge-tts (voice selection, rate/volume/pitch), matching narration length to video with atempo/apad, and multi-scene pacing with breathing room. Targets moviepy 2.x and Python 3.x on macOS, Linux, and Windows.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

Analyzes and automatically optimizes existing agents by improving system prompts and tool configuration.

Creates new agents from user requirements by generating Python implementation and mcp_config.

A professional skill for App Evaluation (evaluating app's performance with score) and App Improvement (giving professional suggestions for improving the app's performance).

HTML 转图片 skill - 将 HTML 文件或内容通过 agent-browser 渲染并截图为图片。适用于生成信息图、社交媒体配图、数据可视化截图等场景。

小红书发布 skill - 通过 agent-browser (CDP) 自动发布小红书图文笔记，支持多图上传、标题正文填写、一键发布。使用场景：自动化发布图文笔记到小红书创作中心。

小红书搜索抓取 skill - 通过 agent-browser (CDP) 抓取小红书搜索结果，支持列表+详情、多格式输出。使用场景：按关键词抓取笔记列表与正文、生成 RSS/JSON/Markdown。

This skill is used for segmented reading and organization when facing large-scale knowledge bases or web pages. It captures original content segment by segment, summarizes key points in real-time, and continuously deposits them into the knowledge base, ensuring orderly information ingestion, clear structure, and traceability.