sheng1111
- 20 skills
- 0 followers
- 15 hours ago last updated
- ▌ Obsidian Inbox Organizer 2 · sheng1111Safely initialize or organize an Obsidian vault by turning loose captures into traceable Sources, reusable Knowledge, Projects, work reviews, concrete tool application cases, and ready-to-paste System Prompts. Use when the user asks to initialize or organize an Inbox, vault, sources, knowledge, or fragmented notes.
- ▌ Obsidian Inbox Organizer · sheng1111 bundleSafely initialize this workflow in an Obsidian vault, incrementally distill loose captures into connected knowledge, or refactor existing Sources and Knowledge when explicitly requested. Handles URLs, images, attachments, tools and repositories, verification, deduplication, canonical notes, meaningful links, noise removal, work reviews, concrete application cases, and reusable Knowledge System Prompts. Use when the user asks to initialize or organize an Inbox, vault, knowledge base, work review, or fragmented captures.
- ▌ Memory · sheng1111Task memory system for recording observations, findings, and learnings during task execution. Short-term memory per task, long-term memory across tasks.
- ▌ Patrol · sheng1111Patrol social media for a keyword like a coast guard hunting targets. Actively search, monitor posts, and collect opinions about a topic. Use when user wants to know what people are saying about something on social media.
- ▌ App Action · sheng1111Quick single-step app operations like launch, tap, type, swipe. Use for simple actions. For research, browsing, or multi-step tasks, use app-explore skill instead.
- ▌ App Explore · sheng1111 bundleOperate Android apps via MCP tools with a research mindset. Browse multiple sources, read content thoroughly, and form conclusions based on evidence. Use for any task requiring app exploration, research, or multi-step operations.
- ▌ Comment Post · sheng1111Post a pre-composed comment to a post/article and verify publication. Does NOT generate comment content (use comment-draft skill first). Use after user approves a draft comment.
- ▌ Comment Scan · sheng1111Deep-read and analyze comments on a post/article. Identify consensus, controversies, common questions, and notable quotes. Use when user wants to understand public opinion on a specific post.
- ▌ Device Check · sheng1111Check device connection, ADB status, and device info. Use before starting automation, when connection issues occur, or when user asks about device status.
- ▌ Social Media · sheng1111 bundleSocial media and dating app operations including like, comment, share, send message, follow, and swipe matching. Supports LINE, Facebook, Instagram, Threads, WeChat, X, Telegram, YouTube, Gmail, and dating apps (Tantan, Tinder, Bumble, etc). Adapts to different UI languages and version differences.
- ▌ Task Clarify · sheng1111Convert ambiguous user requests into structured task specifications with clear parameters and success criteria. Use at the start of complex tasks to ensure requirements are understood.
- ▌ Troubleshoot · sheng1111Diagnose and fix MobileAgent issues systematically. Use when encountering ADB problems, Unicode failures, MCP errors, device connection issues, or automation failures.
- ▌ Comment Draft · sheng1111Generate 1-3 comment drafts based on article context, tone, and identity constraints. Does NOT post comments or execute app operations. Use when user wants to compose a comment before posting.
- ▌ Search Triage · sheng1111Multi-round search with keyword variants on social platforms. Browse results, open promising items, check engagement, and produce a ranked candidate list. Use when user needs comprehensive search results.
- ▌ Unicode Setup · sheng1111Setup Unicode text input for Chinese, Japanese, Korean, and emoji. Install DeviceKit for MCP tools or ADBKeyboard for Python scripts. Use when text input fails or shows garbled characters.
- ▌ Wechat Browse · sheng1111Browse specific WeChat Official Accounts (公众号) by name, navigate to their article history, and extract articles. Use when user wants to read content from specific WeChat accounts.
- ▌ Screen Analyze · sheng1111Analyze current screen state, find UI elements, and get precise coordinates for automation. Use before performing actions or when user asks about what's on screen.
- ▌ Article Extract · sheng1111Extract complete article content including full text, metadata, and expanded sections. Scrolls to load all content, expands "read more" sections. Use when user needs full article text, not summaries.
- ▌ Content Extract · sheng1111Extract complete article content (not summaries) with structured NLP analysis. Identifies who, what, when, where, and objects. Outputs JSON for API integration. Use when user wants full text extraction with analysis.
- ▌ Progress Report · sheng1111Report task progress with evidence including screenshots, text excerpts, and data summaries. Ensures claims are backed by proof. Use periodically during long tasks or when user asks for status.