Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
Productivity
Guo Degang's thinking framework and way of expression. Based on deep research across 9 dimensions (including original-text analysis of 3 first-hand works plus 6 dimensions of web research) and 80+ first- and second-hand sources, distilled into 6 core mental models, 10 decision heuristics, and a complete expression DNA. Purpose: to serve as a thinking advisor, analyzing problems, reviewing decisions, and providing feedback from Guo Degang's perspective. Use when the user mentions "from Guo Degang's perspective", "what would Guo Degang think", "Guo Degang mode", or "guodegang perspective". It should also trigger even if the user only says "help me think from Lao Guo's angle", "what would Guo Degang do", or "switch to Guo Degang".
Use when mentioning Chaos courses, course study, methodologies, mental models, course search, or distilling course transcripts; when extracting value points / sparking ideas / summarizing selling points / pricing products & services / uncovering startup opportunities; or when asked general business decision questions.
Activate when the user needs to understand "why people do this" (especially seemingly irrational behavior), design management systems or incentive mechanisms, assess the credibility of a professional advisor's recommendations, or try to persuade others to change their behavior. Typical trigger signals: the user is puzzled by "why behavior didn't change even though the system did"; the recommendations of the user's advisor, subordinate, or partner happen to be especially favorable to themselves; the user needs to design a system that makes others act as desired. Not applicable to purely technical solution design or scenarios where human behavior need not be considered.
A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).
Search WeChat official account articles by keyword and return titles, summaries, timestamps, account names and links. Use this skill when the user says things like: search WeChat official accounts for articles about AI, find a few posts about personal finance, or any similar WeChat official account article search.
Compile documents, notes, web content, transcripts, research materials, or code repositories into an evidence-grounded, visualization-ready knowledge graph with semantically typed entity nodes, statement-level provenance, and typed relationship edges. Use with ov compile to create or incrementally refresh `entities/*.md` node artifacts and a root `relations.jsonl` edge file for people, organizations, groups, animals, places, products, projects, systems, services, modules, datasets, standards, documents, events, and other identifiable things.
PRD expert advisor mode. Three well-known PM/CEO personas act as advisors, proactively review the PRD, identify weak points and give improvement plans that can be written directly into the document; the user accepts/rejects to decide whether to adopt them. Triggers: /li-prd-review, 「强化 PRD」「专家顾问」「帮我看看这份 PRD」「让专家来审 PRD」「PRD 有哪些漏洞」「帮我优化 PRD」「PRD 还有什么问题」. Even if the user does not say "PRD", trigger this skill whenever they want expert personas to review a product document and give improvement plans. DO NOT trigger for: PRD generation tasks (use li-prd), content script generation, topic analysis and other content creation tasks.
Use this skill when the user says 「获取逐字稿」「转录视频」「提取文稿」「帮我扒一下」「这个视频说了什么」「这个博主说了啥」, or gives a video link and asks to extract its content. Even if the user just drops a video link without saying what for, trigger it proactively as long as the context involves benchmark analysis or content extraction. It calls video2text.py to get the raw transcript, an AI proofreads common speech recognition errors, and after identifying the author it archives to the benchmark blogger directory. Do not trigger for: analyzing viral patterns (use li-analyzer), recording your own topic ideas (use li-recorder), writing your own scripts (use li-writer). Use when the user wants to "get transcript", "transcribe video", "extract script", or gives a video link for content extraction. Runs speech-to-text, AI proofreads, and archives to benchmark blogger directory.
Compile heterogeneous knowledge sources—including documents, notes, web content, transcripts, research materials, and code repositories—into a Karpathy-style, evidence-grounded LLM Wiki with a maintained index; default entity and concept pages; and selective method, comparison, analysis, or instruction-requested summary pages. Use with ov compile to create or incrementally refresh knowledge that is easy for people and agents to retrieve, navigate, and reuse.
Activate when the user needs to understand extreme events (bubbles, crashes, mass hysteria, cults, mob behavior), diagnose systemic organizational failure, or assess the risk of multiple psychological, market, or institutional forces stacking in the same direction at once. Typical trigger signals: the phenomenon the user describes is "far beyond what any single factor could explain"; the user tries to explain an extreme outcome with a single cause; the user worries that "multiple adverse factors will erupt at the same time." Not applicable to routine single-factor decision analysis or gradual incremental change assessment.
Memory recording feature of the AI personal memory system. Automatically distills key information from conversations and records it at the appropriate level. Use cases: (1) when the user says "record this in the memory system", "remember this", "write down this conversation"; (2) when important events, decisions, or expressions of preference are detected; (3) when the user completes an important task or makes a decision. The skill automatically determines whether to record it at the L1 situational layer, L2 behavioral layer, or L3 cognitive layer, or to suggest updating the L4 core layer.
Identify which region of the mirror world a current event falls into (control domain / blind domain / probability domain / risk domain), adopt the corresponding survival strategy, and avoid facing the uncertainty problems of the right-side world with the deterministic rules of the left-side world. Applies to strategy-making for investment decisions, startup choices, and career transitions, as well as postmortem analysis of "why did it fail despite being executed as planned." Not applicable to pure skill improvement and deliberate practice scenarios (those naturally belong to the control domain). Key trigger signals: "why doesn't it work even though I tried hard," "does this depend on skill or luck," "how should I deal with uncertainty," "how do I increase my success rate."
Create scarcity by combining multiple abilities, turning personal value from "replaceable" to "unique." Applies to career planning, personal brand positioning, and core competitiveness design. Use this skill when you feel "I know a bit of everything but master none," or when you want to build a distinctive moat in competition. Not applicable to stages where no single standout ability has been established yet (single-point value comes before combination), nor to scenarios requiring deep specialization in a single field (such as academic research). Key trigger signals: "I have no particular advantage," "others can do what I do," "how do I make myself irreplaceable," "I have broad interests but can't make money."
Searches across your Notion workspace, synthesizes findings from multiple pages, and creates comprehensive research documentation saved as new Notion pages. Turns scattered information into structured reports with proper citations and actionable insights.
Use Octen's search tools for anything that needs live information from the web. Make sure to use this skill whenever the user wants to look something up, check current or recent information, find sources, compare products / vendors / pricing, research a topic, or read a specific web page — even if they never mention Octen or the word "search". Also use it before falling back to any built-in web search tool.
View, create, update, and delete Outlook or Microsoft 365 calendar events with the cola-outlook-calendar CLI. Use for Outlook Calendar schedules, availability, and event changes, or when the user says "outlook 日历"、"看看我 outlook 上的安排"、"帮我在 outlook 上约个会".
Brainstorm a research idea and decide what a new paper should cover — generate candidate framings, pick the central contribution, and draw the in-scope vs. deferred boundary. Use whenever the user has a rough interest, preliminary results, or a problem and asks "what should this paper be about", "help me brainstorm an idea", "what should I include / cover", "is this one paper or two", "help me scope this", "what's the contribution here", or "I have an idea but don't know how to frame it". Default to running this upstream of research-idea-stress-test (which vets one chosen idea) and paper-argument-planner (which structures a paper once the contribution is fixed). Always check first whether the input is sufficient (a seed, the field, the motivation); if not, request the missing context before brainstorming rather than inventing a direction. Hands off per-claim evidence to benchmark-and-baseline-selector, novelty grounding to literature-triangulation, and section logic to paper-argument-planner.
Control Android cloud phones via the `pb` CLI. Make sure to use this skill whenever the user mentions logging into apps, installing apps, browsing on a phone, opening an app (Twitter, Google Play, WeChat, Instagram, TikTok, WhatsApp, Telegram, etc.), searching on phone, checking what's on screen, taking screenshots, tapping buttons, typing text, swiping, or any task that involves an Android device. This skill applies even if the user doesn't say "phone" or "pb" — phrases like "help me log in to Twitter", "install WeChat", "open Google Play", "search for something", or "check the screen" all imply phone operation. Always prefer pb over desktop browsers or Playwright for mobile and app tasks.
Identify and leverage potential-energy gaps (efficiency / scale / cognition / gravity) to choose a high-potential position that amplifies the competitiveness of an individual or organization. Applies to career choices, startup direction decisions, and industry trend analysis. Use this skill when you face a decision about "which platform/company/industry to choose," or want to understand why equal effort yields vastly different results. Not applicable to the execution stage where the track is already locked in and only hard work remains (at that point, return to improving specific abilities), nor to purely internal management optimization (potential energy is an external lever). Key trigger signals: "why doesn't it work even though I try so hard," "industry A or industry B," "does this industry still have a future," "how do I ride the momentum."
Pushover from the terminal: send alerts, watch emergency receipts, inspect quota, and keep a redacted local... Trigger phrases: `send me a pushover`, `notify me on my phone`, `check pushover quota`, `watch pushover emergency receipt`, `validate pushover user key`, `use pushover-pp-cli`.
Use whenever you need to contact someone by email and get their answer back, especially to unblock yourself. Proactively, usually without asking the user first: you are stuck on a tool, library, API, or service and its docs, README, footer, or llms.txt list a help@, support@, dev@, docs@, or agent@ address; you hit a bug worth reporting to the vendor; you have a question only the product owner or maintainer can answer; you want to ask a vendor's support or dev agent something the way you would grep their docs. Reactively: the user gave you a mailto link or contact address, or asked you to email someone or “ask them” and report back. Sends your message and waits for the threaded reply in one step (surface sales@, billing@, or account@ addresses to the user first). No SMTP, no API keys. To receive mail a service sends you, or answer mail someone sent you, use the primitive-inbox skill. Part of the Primitive CLI (primitivedotdev, primitive.dev; the `primitive` or `prim` command)
Use whenever you need an email address to receive mail, read or wait for it, or answer it. Proactively, usually without asking first: a service is about to send a verification code, OTP, one-time or 2FA code, confirmation or magic link, or password reset and you need to read it; you must confirm an email address to sign up for a service; you need a fresh, managed, throwaway, or burner address; you sent something and need to see what lands. Reactively: the user asks “did it arrive?”, “check the inbox”, “wait for the email”, “answer that email”, or wants an address for replies, codes, receipts, or alerts. Provides a managed `*.primitive.email` address, `primitive emails latest` and `primitive emails wait` to read and wait for mail, `primitive reply` to answer an inbound email in its thread, and hosted Functions for inbound mail. To start your own conversation and wait for the answer, use the primitive-chat skill. Part of the Primitive CLI (primitivedotdev, primitive.dev; the `primitive` or `prim` command)
Senior product manager for end-to-end product development from requirements to design. Use when users need: (1) Requirements analysis and product direction, (2) User personas and scenario analysis, (3) Feature and interaction design, (4) PRD writing, (5) Feasibility assessment. Triggers: product design, requirements, user persona, PRD, feature planning, MVP, user story.
Generates structured project status reports from monday.com - a customizable AI work platform with dedicated products for work management, CRM, dev, service, and campaigns. Surfaces overdue items, blockers, workload distribution, and recent activity across one or more boards. Use when user asks for "project status", "board summary", "what's overdue", "status report", "how's the project going", "weekly update", "board overview", "progress update", or "team workload".
Business registration information and personnel relationship lookup (look up companies and owners). Use when: the user needs to look up companies, businesses, owners, shareholders, executives, legal representatives, outbound investments, personnel appointments, business registration information, entity identification, etc.
Research a factual, codebase, product, or external-documentation question using primary sources and return an evidence-backed answer. Use when the user asks to investigate, compare docs, find current behavior, answer a research ticket, or gather facts before a spec, prototype, or implementation.
Turn a bibliography or literature corpus into a knowledge graph of papers, authors, venues, topics and citation links, then analyze it (citation clusters, key papers, bridging work, research gaps) or export it to Neo4j, Gephi, NetworkX, Obsidian or Markdown-graph tools such as graphify. Use when mapping a research field, building a citation network, finding influential or bridging papers, or visualizing how a literature connects. Zero-dependency builder with optional OpenAlex enrichment.
Use when the user explicitly asks to "extract a review topic from a file/image/webpage/description", "generate a structured output of topic + keywords + core questions", or requests the old name get-review-theme skill. Supports multiple input sources such as files (PDF/Word/Markdown/Tex), folders, images, natural language descriptions, and webpage URLs; automatically identifies the input type and extracts content, producing structured output directly usable by research-literature-review and other literature review skills.
Summarizes an RHDH sprint that has finished or is finishing, from Jira RHIDP, RHDHPLAN, RHDHBUGS, and RHDHSUPP: committed versus completed story points, scope added mid-sprint, per-member breakdown, epic progress, the demo checklist with RHDH file and slide naming, and the velocity trend. Use for "sprint report", "what did we complete this sprint", "sprint review prep", "how did the sprint go", or "which demos do we owe". Looks back at work already done; preparing the sprint that has not started yet is a different job, and a single issue such as RHIDP-1234 is not a sprint.
Scores features using both RICE and strategic alignment for nuanced prioritisation. Use when asked to prioritise features, build a priority matrix, combine quantitative scoring with strategic fit, or decide what to build next with multiple competing initiatives. Produces a scored priority matrix with RICE scores, strategic alignment ratings, quadrant placement, and sequencing recommendations.
Create structured roadmap presentations calibrated to any audience. Use when asked to build a product roadmap, present roadmap to leadership, create a roadmap slide, or communicate quarterly plans to execs, teams, or customers. Produces an audience-calibrated Now/Next/Later roadmap with strategic context, initiative tables, success metrics, and explicit deprioritisation rationale.
Transform learning content (like YouTube transcripts, articles, tutorials) into actionable implementation plans using the Ship-Learn-Next framework. Use when user wants to turn advice, lessons, or educational content into concrete action steps, reps, or a learning quest.
Organizes and rewrites prose and notes into outline-style bullet points built around claims and supporting evidence. Use for requests such as "write an outline" or "organize the key points into bullets". Do not use for listing items or task management. For slide creation, do not launch it on its own; it handles only the outline structure as a dependency of the `shunk031-writing-technical-slides` skill.
Reorganizes or rewrites articles and notes as concise telegraphic-style outline bullets, omitting words clear from context and conveying one idea per line. Use for requests like “write a report,” “make it an HTML report,” “in telegraphic style,” or “shorten the bullet points.” Do not use when the request is only to proofread or polish a report that remains in paragraph form. For technical slides, do not trigger independently; it only handles the writing style as a dependency of the `shunk031-writing-technical-slides` skill. Do not use to create slide files
Use when the user explicitly asks to build, extend, validate, repair, rebuild, or merge an LLM-optimized knowledge wiki from markdown notes, documentation, source code, or mixed folders. Default output is a single stable sibling `<source>.wiki/` with full history in a private git repo under `.llmwiki/git/`; `--layout-mode in-place` transforms the source folder itself, and `--layout-mode hosted --target <path>` honours a user-provided `.llmwiki.layout.yaml` contract. SKILL.md is the entry point; detailed operation instructions are loaded on demand from `guide/` per the routing procedure below.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.