agentbay-ai/agentbay-skills
15 skills3.1k installs705 starsGitHub
Install
npx skills add https://github.com/agentbay-ai/agentbay-skillsSkills in this repo
1Stock Watcherstock-watcher from agentbay-ai/agentbay-skills standardizes Chinese A-share watchlist management for Clawdbot agents. It supports adding stocks by 6-digit codes, viewing formatted watchlists, removing individual tickers, clearing the entire list, and fetching performance summaries for all watched symbols. Installation scripts create a standardized watchlist directory under the user home path to avoid path confusion across sessions. Developers automating equity monitoring workflows reach for stock-watcher when they need repeatable CLI-style watchlist commands instead of ad hoc spreadsheet tracking.1kinstalls2Weibo Hot Searchweibo-hot-search is a browser-automation skill from agentbay-ai/agentbay-skills that queries Weibo hot search leaderboards—including entertainment category topics ranked by heat score—and returns top results as Markdown. It depends on `wuying-agentbay-sdk` (`python3 -m pip install wuying-agentbay-sdk`) and runs `python3 scripts/browser-use.py` with step-by-step navigation instructions to weibo.com. Developers reach for weibo-hot-search when building trend monitors, content calendars, or social listening agents focused on Chinese microblogging data. Output includes topic name, heat score, and rank for entries meeting configurable heat thresholds such as ≥50000.514installs3Boss Job Searchboss-job-search in agentbay-ai/agentbay-skills drives BOSS Zhipin (zhipin.com) job research through the wuying-agentbay-sdk and a bundled browser-use.py script. Agents execute natural-language step lists to search keywords such as AI Infra or Python engineer, apply filters like company size (10,000+ employees) or salary bands, and extract up to 15 listings per run with title, company, size, salary, location, publish date, and summary. Output is markdown tables plus optional aggregate stats (count, average and min/max salary). The skill documents three worked examples, captcha and timeout handling, and a 1–2 minute runtime expectation per search. Developers reach for boss-job-search when evaluating China's tech job market, comparing compensation bands, or building agent workflows that need structured job data without manual browsing. Triggers include requests to search Boss Zhipin, filter by company scale, or pull latest role requirements.297installs4Web ScraperAgentBay web-scraper skill teaches Claude Code how to build reliable site scrapers for agent workflows: request pages, parse DOM or APIs, handle pagination, and return clean structured data for automation and research pipelines.211installs5Find SkillsHelps users locate and evaluate agent skills from catalogs and community repos: interpret goals like 'how do I do X', surface matching SKILL.md packages, compare alternatives, and guide installation so teams adopt proven instructions instead of building from scratch.184installs6Douban Movie ReviewAutomates Douban movie-review collection and analysis for agent workflows: query titles, scrape or fetch ratings and comments, summarize sentiment, and return structured review data for content or recommendation features.160installs7Qwen Wanx Comic Genqwen-wanx-comic-gen is an OpenClaw-oriented agent skill backed by a Python generator that turns natural-language prompts into manga and comic-style images through Alibaba Cloud DashScope Tongyi Wanxiang (wan2.6-t2i). Solo builders and indie teams use it when they want illustrated assets—four-panel gags, Q-style characters, or branded mascots—without leaving their coding agent workflow. The script creates an asynchronous generation job, polls until completion, saves files locally, and surfaces results as MEDIA lines the agent can ship in chat or downstream automations. Configuration expects a DashScope-capable API key on the primary OpenClaw model provider entry, with environment-variable fallback when the config uses placeholders. Because the model omits a dedicated style knob, effective prompts must spell out anime, panel layout, and character tone explicitly. It fits builders already on OpenClaw who need repeatable comic art generation rather than a general-purpose design suite.117installs8Wuying Browser Usewuying-browser-use is a Claude Code skill that automates browser interaction through the Wuying AgentBay SDK. A developer invokes it to navigate sites, fill and submit forms, take screenshots, and extract data by describing the task in Chinese or English. Each command runs independently without keeping session state and takes one to two minutes to complete.107installs9Qwen Imageqwen-image is a Claude Code skill that generates images from text prompts using Alibaba Cloud's Qwen Image API (通义万相 / DashScope). It runs a bundled generate_image.py script via uv, supports Chinese and English prompts, and lets the user pick sizes and models. By default it returns an image URL (parsed from a MEDIA_URL line) to render in Markdown and downloads only when asked. A developer invokes it for text-to-image generation, and it needs a DashScope API key.105installs10Amap Trafficamap-traffic is a Claude Code skill that queries real-time road conditions and plans optimal driving routes using AMap (高德地图) traffic-status and direction APIs. A developer invokes it to check congestion on a road or area and to compute the fastest self-driving route with strategy=2 real-time-traffic weighting. It reads the AMAP_KEY dynamically from openclaw.json on every call so an updated key takes effect immediately. It targets webchat, DingTalk, and QQ channels.94installs11Im Reminderim-reminder is a Claude Code skill that schedules one-time and recurring reminders that wake the agent at a set time to message the user. It supports at (one-shot timestamp), every (interval), and cron schedule kinds, and auto-detects the current IM channel and user ID so the reminder is delivered to the original channel. A developer invokes it for requests like remind me in 5 minutes or every hour. It requires fixed job fields (enabled, isolated session, deliver true) to avoid missed deliveries.81installs12Agentbay Aio Skillsagentbay-aio-skills is a Claude Code skill that runs user code in an isolated AgentBay cloud sandbox instead of the local terminal. It creates a code_latest session via the wuying-agentbay-sdk, executes code with run_code, and returns the result. A developer invokes it for any run or execute-code request in Python, JavaScript, R, or Java, including plotting. It reads an AgentBay API key from a config file or environment variable and has a 60-second per-execution timeout.61installs13Agentbay Monitor Skillsagentbay-monitor-skills is a Claude Code skill that runs a public-opinion monitoring pipeline ending in a sentiment report. It crawls a platform (Baidu, Bing, Xiaohongshu, Weibo, Douyin, or Zhihu) by keyword, the main agent scores sentiment per item using a bundled instruction file, and report.py generates a Markdown/JSON report with optional PDF. A developer invokes it when asked how sentiment or opinion looks on a topic. It is nearly identical to agentbay-monitor-skill and requires an AGENTBAY_API_KEY and platform login for social sources.60installs14Moltbook Hot Postsmoltbook-hot-posts is a Claude Code skill that retrieves hot and latest posts from Moltbook, an AI-agent community. It drives a bundled browser-use.py script to visit moltbook.com, set the time and Top/New filters, open the top post, and return a Markdown summary of its content and interaction stats. A developer invokes it to check trending topics or the latest discussions in the AI-agent community. It reads the main post and the first few replies rather than every comment.51installs15China Stock Analysischina-stock-analysis is a Claude Code skill for value-investing analysis of China A-share stocks, aimed at low-frequency retail investors. It screens stocks by valuation, profitability, growth, dividend, and financial-safety criteria, runs deep single-stock financial analysis, compares peers within an industry, and computes intrinsic value via DCF, DDM, and relative valuation. It uses akshare for public financial data and produces a Chinese analysis report with scoring, risk flags, and a margin-of-safety verdict. A developer invokes it to screen, analyze, compare, or value A-shares.2installs