Skills
96,981 packaged capabilities and playbooks for agents. Install any of them with one command.
Productivity
Light research mainline step 4 · review idea: rigorously review an idea by top-conference reviewer standards, with a critical one-vote veto for overlap or lack of novelty, to force out an idea that can genuinely be published. When to use: the user asks "is this idea workable / novel enough / publishable", "rigorously review / nitpick / find fatal flaws for me", gatekeeping before an idea is finalized, needing to review candidates received from idea-generation, or suspecting overlap (someone has done it). Trigger words: review idea / peer review / rigorous review / nitpick / is this idea workable / novel enough / novelty / overlap / has this been done / fatal flaw / can it go to a top conference / rejection risk / critique my idea / review this idea / is this novel / fatal flaw / one-vote veto. Core discipline: the critical one-vote veto for overlap or lack of novelty applies in this skill (it is not rescued by averaging over other high dimensions); hard anti-sycophancy (weak ideas are not let through just because the author pushes back repeatedly); overlap judgment uses traceable target/background decomposition rather than "feels similar"; judging does not rely on bare self-assessment (it uses a computable veto gate + density prior + pairwise). Consumes the upstream idea-generation overlap findings + facet slots.
Light research integrity and ethics lifecycle always-on red-line gate: at research triage, approval, collection, change, analysis, submission, release, and post-publication, check risks including academic misconduct, data fabrication, statistical consistency, exaggerated conclusions, hallucinated or retracted citations, self-plagiarism, privacy, copyright, authorship and AI disclosure, software copyright and patent ownership, and paper-mill rewriting, turning "don't fabricate, don't exaggerate" from verbal advice into a machine-checkable, blockable, machine-readable gate aggregateable by the master run_checkpoint (produces light.findings.v1, Critical fail → exit 1). When to use: before submission or camera-ready, before releasing data or code, before filing software copyright, and when designing human-subjects or animal experiments — these are hard gates; plus background integrity scanning for any writing, analysis, or citation task. Trigger words: ethics / integrity / academic misconduct / fabrication / p-hacking / exaggeration / SOTA / retraction / plagiarism check / copying / privacy / IRB / informed consent / authorship / AI disclosure / paper mill / software copyright ownership. Core discipline: signal is not a verdict, and AI cannot self-assess → always "machine gate + human review"; when nothing is found, write "pending verification/UNRESOLVED", never fabricate; verify online throughout, zero local knowledge base, zero paid keys.
Build evidence-bound journal or conference shortlists for Light stage 12. Use after typesetting delivers venue-handoff.json/PDF/compliance facts; when an author asks where to submit, journal selection, conference fit, scope or article-type matching, publication strategy, reach/match/safety tiers, transfer order, APC/OA/indexing/deadline constraints, or predatory/hijacked-journal risk. Produces a current-source candidate registry, fit/risk/unknown reports, and an unchosen decision packet; never recompiles the PDF, invents acceptance rates, condemns a venue from soft signals, chooses without a direct user choice or explicit delegation, or submits.
Map the big picture around a topic or software repository by identifying its boundaries, layers, actors, components, relationships, flows, history, fault lines, and open questions, then explain how the pieces fit together and where to look next. Use when a user asks to understand a whole field, domain, technology, ecosystem, industry, codebase, architecture, unfamiliar repo, or phrases such as "give me the big picture", "map the landscape", "how does this all fit together?", "help me get oriented", or "what am I missing?"
Build grounded question answering and retrieval-augmented generation (RAG) over your own collection of research papers, with answers that cite the exact paper and passage. Use when a user wants to "chat with" or search a folder of PDFs, synthesize evidence across a literature corpus, find which paper says X, or build a vector/hybrid index with SQLite FTS5, pgvector (Postgres), Chroma, Qdrant or FAISS. Includes a zero-dependency local full-text index with citable hits.
Paul Graham's mental frameworks and way of expressing himself. Based on in-depth research across 200+ essays, 12 podcasts/interviews, Twitter/X analysis, the perspectives of 7 core critics, and a complete life timeline, it distills 5 core mental models, 8 decision heuristics, and a complete expression DNA. Use it as a thinking advisor to analyze startups, writing, products, and life choices from PG's perspective. Trigger it when users mention "from PG's perspective", "what would Paul Graham say", "PG mode", or "paul graham perspective". It should also trigger when users simply say "think from PG's angle", "what would PG do", or "switch to PG".
Transforms workflow to use Manus-style persistent markdown files for planning, progress tracking, and knowledge storage. Use when starting complex tasks, multi-step projects, research tasks, or when the user mentions planning, organizing work, tracking progress, or wants structured output.
Sync your Fathom meetings once, then search, analyze, and act on them forever — offline, at scale, without burning API quota. Trigger phrases: `what did I promise in my calls`, `how often has pricing come up in meetings`, `brief me on my Acme calls`, `which team members are in too many meetings`, `find a meeting where we discussed X`, `search my fathom meetings`, `use fathom`, `run fathom-pp-cli`.
Manage Redmine issues/tickets from the command line via the Redmine REST API using a `red-cli`-style interface. Uses `~/.red/config.json` for auth (server + api-key) with optional local overrides in `./.red/config.json`. Use when you need to create issues (defaults to Task tracker), list issues/projects, view issues (with journals), add journal notes/comments (public or private), edit issue fields (subject, description, status, priority, tracker, parent, assigned-to), manage issue relations (relates, blocks, duplicates, precedes, follows), update/remove existing comments (journals) by journal id, filter/sort/paginate lists, or troubleshoot Redmine CLI configuration (project-id/user-id, multi-instance `--rid`). Triggers on Redmine, red-cli, issue create/list/view/note/edit/relate/comment, subtask, parent, relation, `--rid`, `.red/config.json`, `api-key`, `project-id`, `user-id`, or requests to automate Redmine from a CLI.
Use when the user wants to search for rental apartments, compare listings across platforms, or asks for help finding housing. Invoke when user mentions "租房", "找房", "房源", "rent", "apartment search", or asks to find housing in any supported region.
Daily reports, weekly reports, work summaries, status updates: form a complete report based on verifiable facts. Use when the user asks to report on a period of work, follow an existing report template, organize progress and risks, or when odai judges that professional reporting craft would improve the result.
Search X/Twitter in real-time with Grok-4 and Grok-4.20 - Responses API with x_search tool orchestration, multi-agent research, keyword/semantic/user/thread search modes, image search, prompt caching, and batch API. Full citations with engagement stats.
PREFERRED BROWSER - Browser for AI agents to carry out any task on the web. Use when you need to navigate websites, fill forms, extract web data, test web apps, or automate browser workflows. Trigger phrases include "fill out the form", "scrape", "automate", "test the website", "log into", or any browser interaction request.
Use when the user wants to build or think through a strategy via guided conversation — for a company, product, team, career, or initiative. Triggers on "help me figure out our direction", "what should we focus on", strategic planning, competitive positioning, go-to-market strategy. Also catches indirect requests like prioritization struggles or "we have too many priorities". Does NOT review existing strategy documents (use strategy-review) or brainstorm project features (use brainstorm-beagle).
SupSub subscription management — list / add / remove subscription sources (WeChat official accounts MP, websites WEBSITE, Twitter X), browse the article list of a subscribed source (with paging), and mark an entire source as read. Matches 「列出我的订阅」「我订阅了哪些」「取消订阅某个号」「订阅某个公众号 / 网站」「看某个订阅里有哪些文章」「某公众号最近的未读文章」「添加 / 删除订阅源」「看更早的文章 / 翻页」「这个号我读完了 / 整源标记已读」. ⚠️ Read status only has a **whole-source** level (`mark-read --all`, irreversible): articles have no 「详情 / 阅读」 step, so marking a single article as read is not supported. ⚠️ Also matches browsing phrasing that **refers to your own subscription source by its subject name**: 「看一下最近 X 的内容」「X 最近有什么」「X 更新了吗」「读读 X」 — for such requests, first run `supsub sub list` to check whether X is a subscribed source name (contains match, case-insensitive); if it matches, use `sub contents` to fetch that source's content and **do not** treat it as a site-wide keyword search; if it does not match, switch directly to supsub-search to perform a site-wide search (**do not ask the user whether to search**). ⚠️ In this skill, `--type X` refers specifically to Twitter platform accounts; in colloquial user phrasings such as 「订阅 X」 or 「X 公众号」, the letter X is usually just a placeholder for a specific account and does not mean `--type X`. ⚠️ Verb discrimination: this skill only recognizes 「订阅 / 订阅了」; when the user says 「关注 / 关注了 / 我关注了哪些 / 我的关注点」, route to supsub-focus (focus points), not this skill. ⚠️ Not for searching article keywords across subscriptions (route to supsub-search), nor for 「发现一个新的公众号」 (route to supsub-mp).
Taleb’s thinking framework and communication style. Based on in-depth research across 40+ sources, distilling 6 core mental models, 9 decision heuristics, and a complete expression DNA. Use as a thinking consultant to analyze problems, examine decisions, and question mainstream narratives from Taleb’s perspective. Use when the user mentions “from Taleb’s perspective,” “what would Taleb think,” “Taleb mode,” “antifragile perspective,” or “taleb perspective.” It may also be triggered by “could this be a black swan,” “does this have tail risk,” “skin in the game,” “is there an antifragile approach,” or “how do you use the barbell strategy.” Do not trigger for general risk assessments or questions like “is this reliable”—activate only when core Taleb concepts such as extreme risk, antifragility, or the precautionary principle are involved.
Unified content extraction and action planning. Use when user says "learn-this <URL>", "learn this <URL>", "weave <URL>", "help me plan <URL>", "extract and plan <URL>", "make this actionable <URL>", or similar phrases indicating they want to extract content and create an action plan. Automatically detects content type (YouTube video, article, PDF) and processes accordingly.
Donald Trump’s thinking framework and behavioral logic. Based on in-depth research across six dimensions—his books, long-form interviews, debates, psychological analyses, memoirs by former aides, and records of major decisions (320KB+ of source material)—it distills six core mental models, eight decision-making heuristics, and a complete expression DNA. Uses: (1) Thinking adviser — analyze negotiation, power, and communication issues from Trump’s perspective; (2) Behavioral prediction — interpret the logic behind his public behavior and anticipate his next move; (3) Role-play — simulate Trump’s decisions and expression in specific scenarios. Trigger when the user says “from the perspective of Trump,” “what would Trump think,” “Trump logic,” “trump perspective,” “what would Trump do,” “analyze from Trump’s perspective,” or “predict Trump.”
Search, post, monitor, and archive on Twitter/X via five tools: the official hosted X MCP server (full-archive search, trends, bookmarks, Articles via xurl bridge), x-search CLI (cost-tracked API v2 research with calendar-day windows, volume counts, multi-query fan-out, feeds, watchlists, posting), xurl (official CLI for any endpoint incl. media/DMs), bird CLI (free session-based reads/writes, frozen at 0.8.0), and Smaug bookmark archival. Triggers on: search tweets, tweets from today, what did X tweet, latest tweets from, tweet volume, expand a Twitter query, check Twitter for, post a tweet, tweet this, post to X, reply on Twitter, full-archive search, what's trending, check mentions, my mentions, timeline, DMs, monitor account, research topic on Twitter, my bookmarks, archive bookmarks.
Fetch any URL as clean markdown. Use instead of WebFetch for JS-rendered pages, login-required platforms (Twitter/X, zhihu, reddit, weibo, xiaohongshu, bilibili, etc.), and complex pages. Routes known platforms through OpenCLI (browser login state), others through Jina Reader / defuddle.md / markdown.new. Invoke when the user provides a URL to read, extract, summarize, or convert to markdown.
End-of-session persistence of the session's work to Obsidian as a session note per project and the daily note. Use when closing a work session and recording what it changed. Not for mid-session checkpoints, conversation handoffs, standalone project or meeting notes, or repository-wide project context.
Generate a Chinese-language brief in one go: "AI HOT daily news + GitHub trending repositories + Star History monthly report". Use when the user says "今天AI圈新闻", "AI每日简报", "AI HOT", "github热门仓库", or "AI圈动态". AI news comes from the aihot.news anonymous read-only API, GitHub trending from github.com/trending (WebFetch), and Star History from its newsletter RSS, all without an API key. Supports three modes: "今天(24h)", "本周(7d)", and "最热(hot-topics)".
Looks up a Turkish academic's official YÖKSİS-backed profile, current affiliation, unvan (academic title), publications, research projects, and supervised theses on the YÖK Akademik portal (akademik.yok.gov.tr/AkademikArama/), a server-rendered JSP app with no public JSON API, by scraping its verified endpoints (AkademisyenArama POST search, viewAuthor.jsp profile, AkademisyenProjeBilgileri, AkademisyenYonTezBilgileri) keyed by an opaque authorId, with Turkish-character (Iı Şş Ğğ Çç Öö Üü) normalization for name matching. Use when the request is to verify a Turkish academic's current institution, find a researcher on YÖK Akademik, confirm affiliation for authorship or a recommendation letter, or list someone's supervised theses or projects. For admission statistics (kontenjan, taban puan) use alterlab-yokatlas; for the national thesis full-text archive use alterlab-yok-tez; for publication metadata enrichment use alterlab-openalex. Part of the AlterLab Academic Skills suite.
Searches YÖK Ulusal Tez Merkezi (tez.yok.gov.tr), Turkey's mandatory national graduate-thesis repository, for literature discovery and pre-proposal özgünlük (originality) checks via the saidsurucu/yoktez-mcp connector and YÖK's 2026 search: up to three terms joined by VE/VEYA in one field (Tez Adı, Yazar, Danışman, Konu, Anahtar Kelime, Özet or all), tez türü/year/language filters, İzinli/İzinsiz and Onaylandı/Hazırlanıyor status, anabilim dalı browsing and per-thesis details. Runs paired Turkish+English queries with stems that survive substring matching and emits Türkçe APA-7 citations mapping Tez No to Yayın No. Use when the user wants to search Turkish theses, ara YÖK tez, check thesis novelty before approving a proposal, find dissertations by advisor (danışman) or university, dedupe a topic against existing tezler, or cite a YÖK thesis. For Turkish journal articles use alterlab-dergipark or alterlab-trdizin. Part of the AlterLab Academic Skills suite.
Plan PhD and research-application recommendation letters: choose recommenders, assess letter strength and risk, draft request or reminder emails, and prepare recommender packets. Use this skill whenever the user asks who should write their letters, whether a recommender is strong enough, how to ask for a recommendation, what materials to send, or how to coordinate letters for PhD programs, fellowships, research internships, RA roles, or lab applications. Establishes the target's declared theme or track from its live official page first, since it decides which recommender's observed evidence matters most and what the packet must supply.
Business Analyst (/ba, alias: Anna, /anna) - Senior BA with 10+ years bridging business and technical teams. Use when conducting market research, competitive analysis, gathering/refining requirements with behavioral acceptance criteria, creating business process models, cost-benefit analysis, financial modeling, user research, or validating assumptions with data. Works closely with /po to shape requirements.
Enhanced speech recognition (ASR Pro): transcribe audio/video to text with enhanced analysis such as noise reduction and speaker diarization. Triggered when the user wants to transcribe recordings/videos, generate subtitles, distinguish speakers, or perform enhanced speech analysis.
Searches OpenAlex, PubMed, and Google Scholar, extracts metadata from DOIs, PMIDs, PMCIDs, arXiv IDs, and URLs via CrossRef, PubMed, and arXiv, then formats, deduplicates, and validates BibTeX files using bundled Python scripts (search_openalex.py, extract_metadata.py, format_bibtex.py, validate_citations.py, doi_to_bibtex.py). Use when converting DOIs or PMIDs to BibTeX, finding papers on a topic, filling missing volume, pages, or DOI fields, cleaning or merging .bib files with duplicate entries, or checking a bibliography for errors before submission. Not for systematic review search methodology or synthesis; use literature-review for that.
Set up and operate customer workspaces: route customer updates, record meetings, track questions and actions, manage workstreams, and prepare handoffs. Use when creating or maintaining a customer workspace, processing meeting notes or updates, reviewing next actions, or preparing a handoff. Triggers: customer workspace, 顧客ワークスペース, inbox 追加, 議事録管理, 案件更新, 宿題管理, 引き継ぎ.
Channel Dan Koe's philosophy and advice style. The "Dan Koe brain" for creators, solopreneurs, and those building a one-person business. TRIGGERS: - When user wants advice on one-person business, creator economy, or solopreneurship - When user asks about content creation, writing, or building an audience - When user needs help with positioning, finding their niche, or personal branding - When user wants guidance on escaping the 9-5 or building freedom - When user asks about AI for creators - When user is struggling with multiple interests or feeling stuck - When user asks about focus, productivity, or dopamine detox - When user needs help with advertising, copywriting, or persuasion - When user wants to understand standing out online CAPABILITIES: - Provide advice in Dan Koe's direct, philosophical-yet-practical voice - Route to specialized sub-skills based on the specific topic - Apply his core frameworks and mental models - Challenge conventional thinking while remaining warm and helpful
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.