Recently Added
Fresh skills just added to the collection. Be the first to try them out!
arxiv
by brycewang-stanford
Search and retrieve academic papers from arXiv using their free REST API. No API key needed. Search by keyword, author, category, or ID. Combine with web_extract or the ocr-and-documents skill to read full paper content.
marp-export
by brycewang-stanford
Export Marp slide markdown files to PDF (default), HTML, PPTX, or PNG. Handles theme loading, Chrome/Chromium detection for PDF export, and batch conversion. Trigger when user says "export slides", "convert marp to pdf", "marp export", "slides to pdf", "slides to html", "导出slides", "slides转pdf", "导出演示文稿".
do-agent
by brycewang-stanford
This skill should be used when the user asks to "/do-agent", asks for "Execute complex tasks using multi-agent architecture with context protection", or needs the workflow previously provided by the /do-agent slash command.
China-CF-Study
by brycewang-stanford
根据研究者提供的研究计划书（Research Proposal）执行基于中国制度环境的公司金融类实证研究全流程。启动后第一件事：根据计划书的主题、识别策略、贡献边际与样本范围，从中国-context 英文顶级期刊池（JF/JFE/RFS/JFQA/MS/JCF/JBF/JAR/JAE/TAR/CAR/JIBS/China Economic Review/PBFJ 等 25+ 期刊）中推荐 5 本最匹配的目标期刊（[J1]–[J5]），等待研究者明确选定一本；该期刊决定 main.tex 的 bibliographystyle、Section 骨架、Introduction 风格与表注规范。然后用 Python 完成数据清洗、描述性统计、基准回归、内生性检验（IV/2SLS、DML）、平行趋势、异质性、机制、稳健性检验与图表绘制。LaTeX 表格和图像严格遵循 template/ 示例格式，研究逻辑与排版严格遵循 rule/ 下的《通用实证研究逻辑与规范总结》与《回归表写作规范总结》。数据集与政策集从 asset/ 中按计划书中的关键词检索。当计划书预期的实证结果无法实现时（系数不显著、平行趋势不通过、IV 弱工具、机制不成立等），skill 自动切换备选方案直至完成研究项目。最终交付物：Python 代码 + LaTeX 表格 + 图像（.pdf/.png）。触发条件：研究者提交研究计划书（含 X→Y 假设、识别策略、样本、政策冲击等）。
marp-slides-creator
by brycewang-stanford
专业Marp演示文稿制作助手。支持完整工作流程：工作空间初始化、内容分析、slides制作、多维度审阅、中文语言规范审阅（中文演示文稿）、PNG转换检查、终稿确定。所有产出物集中管理在项目工作文件夹中。当用户提到"制作slides"、"做PPT"、"演示文稿"、"Marp"、"幻灯片"、"presentation"等关键词时自动启用。Professional Marp presentation assistant with complete workflow: workspace initialization, content analysis, slide creation, multi-dimensional review, Chinese language review (for Chinese presentations), PNG conversion check, and finalization. All outputs organized in project workspace.
econfin-idea-finder
by brycewang-stanford
公司金融实证研究的"漏斗式选题查找器"。互动开场先后询问 (1) 研究方向、(2) 候选标题数量 N， 再扫描全球文献（已出版英文学术期刊 + SSRN working paper + 全球高校 department seminar 1 年内日程），基于 Edmans (2024) "1000 Rejections" 红线生成 N 个候选标题，通过并行 subagent（Agent 工具）批量生成计划书 + 查新；每个 subagent 必须强制调用 Skill 工具加载 econfin-proposal 与 novelty-check 两个预设 skill 完成各自模块，只有当 novelty score >= 9 时（即 JF/JFE/RFS 顶刊层次），subagent 才把 proposal + 查新报告合并的 md 写入 F:\Dropbox\CC\选题大全\<研究方向短名>\（以"简短选题名称-分数"命名，子文件夹名由 Step 0 从用户输入的研究方向派生）；< 9 分的选题在 subagent 内部直接丢弃，绝不写盘、绝不输出。当用户说"找选题"、"帮我找选题"、"想做 X 方向"、 "empirical CF idea search"、"批量生成研究计划书"、"100 ideas"、"econfin-idea-finder" 时触发。
Econfin-Proposal
by brycewang-stanford
金融经济学实证论文计划书生成器。根据用户提供的研究方向，生成包含标题、假说、数据来源、实证策略、 预期结果表格、稳健性检验、异质性分析、机制检验等12个完整模块的研究计划书。 内置中国微观/宏观数据库（皮皮侠1599个数据集、马克数据377个数据集）和WRDS国际数据库索引， 自动匹配可用数据源。融合Edmans (2024) "Learnings From 1000 Rejections"的编辑视角作为质量护栏， 确保选题具有真正的边际贡献而非"just another determinant of Y"。 当用户提到以下任何情境时触发：写研究计划书、research proposal、论文开题、选题+计划、 帮我设计一个实证研究、empirical research design、我想研究X对Y的影响怎么做、 帮我找个能发表的选题、generate proposal、写一个可以投稿的研究方案、 研究设计、identification strategy、DID/RDD/IV研究设计。 即使用户只是描述了一个经济金融现象并想知道"能不能做成论文"，也应考虑使用此技能。
master-thesis-review
by brycewang-stanford
Generate master's thesis review reports (硕士论文评阅意见) calibrated to a given score. Outputs academic evaluation and shortcomings/suggestions in Chinese. Trigger when user says "master thesis review" / "硕士论文评阅" / "论文评阅" / "评阅意见" / "thesis review".
figure
by brycewang-stanford
Econometrics skill for generating publication-quality figures for top economics journals. Activates when the user asks about: "econometric figure", "publication figure", "journal figure", "AER figure", "QJE figure", "event study plot", "coefficient plot", "coefplot", "binned scatter", "binscatter", "RDD plot", "parallel trends plot", "kernel density", "distribution plot", "time series plot", "map", "figure formatting", "academic plot", "论文图表", "学术图", "系数图", "事件研究图", "散点图", "分布图", "趋势图", "回归可视化"
md-to-docx
by brycewang-stanford
Convert Markdown files to Word documents (.docx) with proper formatting, Chinese font support (FangSong for all text including headings), black font color, 1.5x line spacing, precise first-line indent (24pt), heading spacing after (1 line), no italic headings, and automatic superscript conversion for citation numbers. Use when converting .md files to .docx, creating Word documents from markdown, or when user mentions Word, DOCX, or document conversion. Requires pandoc.
ml-causal
by brycewang-stanford
Econometrics skill for machine learning methods in causal inference. Activates when the user asks about: "causal forest", "generalized random forest", "GRF", "double machine learning", "DML", "debiased machine learning", "LASSO for variable selection", "post-LASSO", "heterogeneous treatment effects", "CATE", "conditional average treatment effect", "BLP analysis", "CLAN analysis", "causal tree", "honest estimation", "因果森林", "双重机器学习", "异质性处理效应", "条件平均处理效应", "LASSO变量选择", "机器学习因果推断", "去偏机器学习"
berg-pc-design
by qiansui890-wq
Use this skill to generate well-branded interfaces and assets for Berg PC, a Houston trial & business law firm focused on cryptocurrency fraud tracing and recovery — either for production or throwaway prototypes/mocks. Contains essential design guidelines, colors, type, fonts, logo assets, reusable components, and UI kits for prototyping.
scan-to-latex
by mikiway902
Используй этот навык для оцифровки отсканированных технических документов — методичек, конспектов, учебников, курсовых и лабораторных работ, статей — присланных как PDF (включая сканы без текстового слоя и развороты книги по два листа на одном изображении) или как набор PNG/JPG-фотографий страниц. Задача навыка — сохранить ВЕСЬ текст без потерь, перевести ВСЕ формулы в корректный LaTeX, а графики/номограммы приблизить полиномами с указанием диапазона применимости — а не оставлять картинками или искажённым результатом обычного OCR. Применяй, когда просят «оцифровать скан/методичку», «распознать текст со страниц», «перевести формулы в латех», «снять график полиномом», «сделать конспект из фото», «вытащить текст и формулы из PDF», когда документ содержит математику, физику, химию, инженерные расчёты, графики/номограммы или много формул, а также когда часть страниц — нечёткие фотографии тетради или книги. Для документов без формул и без визуальной сложности достаточно обычного pdftotext — этот навык не для них.
investment-analysis
by Hainrixz
Multi-agent investment research and analysis system by Tododeia. Use when the user wants market analysis, investment research, or a summary of current opportunities across crypto, stocks, forex, and commodities. Spawns 5 specialized research agents (4 sector + 1 strategy), adapts to user risk profile, tracks historical accuracy, and generates a branded interactive HTML report served locally. Educational analysis only — not financial advice. Trigger phrases: "investment analysis", "market research", "analyze markets", "investment opportunities", "what should I invest in", "market report", "tododeia", "investment advice", "portfolio recommendations", "run tododeia", "daily market analysis", "weekly report", "análisis de inversión", "analizar mercados", "reporte de mercado", "oportunidades de inversión".
hush
by royashbrook
Use whenever an agent needs to STORE, GENERATE, or USE a secret (API token, key, signing value, password) without ever exposing the plaintext. Replaces the "go set this env var / paste this token into that system" dance with one structured, OS-keychain-backed flow where the value goes straight from source into the consumer and never passes through the agent (no transcript, no logs, no cloud). Two add-paths: a value you GENERATED elsewhere (a vendor token, a PAT) gets pasted in once via a hidden prompt the agent can't see; a value that just needs to be STRONG+RANDOM (an operator key, a webhook signing secret) the agent generates and stores itself. Then it injects straight into the consumer (an env var, or a command's stdin), never printed , so an agent running as the user, with their CLIs already authed, can set server-side secrets and call services without the value ever touching the chat or disk. Triggers: "store this token", "save this key", "add it to the keychain", "generate an operator/signing key", "use the X secret to call Y", or any moment an agent needs a credential to reach a service. macOS, Linux, and Windows backends built in; the never-print contract is portable beyond them.
lobster-feishu-ecommerce
by 23xxCh
飞书龙虾（电商版）— 全宇宙最会用飞书的 AI 数字员工，专为电商团队打造。 当用户发送 "安装飞书龙虾"、"装龙虾"、"lobster-feishu"、"电商飞书助手"、 "https://github.com/23xxCh/onlyyou" 时触发。 一键安装 16 个技能 + 24 个飞书工作流 + 9 个电商案例 + 3 个提示词模板。 安装完成后自动发送说明书。
boar-strategy
by Senpi-ai
BOAR v1.0 — Hard Money vs Paper (Debasement) Hedge Fund. Two books on two wallets, one producer. The LONG "hard-money" book longs scarce real assets — GOLD, BTC (digital gold), SILVER, PLATINUM, PALLADIUM. The SHORT "paper" book shorts the broad market (SP500) + rate-sensitive long-duration growth (RIVN/DKNG/HIMS) — the fiat-denominated financial claims that lose most as the term premium rises under fiscal dominance. The edge is the real-assets-beat-paper-claims spread. HONEST CAVEAT: the loosest hedge of the thesis-fund family — in a pure liquidity melt-up gold AND stocks can rise together; the short tilts to rate-sensitive names to tighten it. Trend-confirmed, conviction-sized. NOT a copy-trader; the runtime owns the LLM gate (pass-through), DSL exits, and all risk.guard_rails. BOAR_LEG env selects the book.
mongoose-strategy
by Senpi-ai
MONGOOSE v1.0 — On-Chain Finance vs Legacy Hedge Fund. Two books on two wallets, one producer. The LONG "on-chain" book longs the on-chain financial rails on Hyperliquid — HYPE (the venue), CRCL (Circle/stablecoins), COIN (Coinbase), HOOD (Robinhood), MSTR + PURRDAT (crypto treasuries) — the disruptors eating legacy finance. The SHORT "legacy" book shorts the incumbents + broad financial-beta (BX/Blackstone + SP500). Trend-confirmed on ABSOLUTE structure, per-name conviction sizing (HYPE/CRCL big; levered treasury proxies small). The edge is the disruptor-vs-incumbent spread. NOTE: the hedge is cross-sector and the short universe is thin, so it leans on the index — a looser hedge than a same-sector pair. NOT a copy-trader; the runtime owns the LLM gate (pass-through), DSL exits, and all risk.guard_rails. MONGOOSE_LEG env selects the book.
eel-strategy
by Senpi-ai
EEL v1.0 — Electrons vs Hydrocarbons Energy-Pair Hedge Fund. Two books on two wallets, one producer. The LONG "power" book longs the AI-power complex on Hyperliquid XYZ — uranium (URNM), gas-fired power (NATGAS), the grid metal (COPPER), on-site fuel cells (BE), rare-earth (USAR) — the electrons AI datacenters need. The SHORT "oil" book shorts crude (Brent + WTI), the legacy transport fuel losing relative ground to electrification. Both legs are ENERGY, so an energy-wide move washes out and the P&L is the electrons-vs-hydrocarbons spread — an energy-sector-neutral pair, not a directional energy bet. Trend-confirmed on ABSOLUTE structure, per-name conviction sizing (uranium/copper big, fuel-cells small). NOT a copy-trader: each book scores its own universe; the runtime owns the LLM gate (pass-through), DSL exits, and all risk.guard_rails. EEL_LEG env selects the book.
sanguo-redeem
by wennyben
Redeem Sanguo game gift codes for configured HK users. Use when the user mentions 三國志,三國,Sanguo,禮包碼,兌換碼,redemption codes, or asks to redeem a code.
skaters code-review skill
by microprediction
uses skaters appropriately, say so and move on — don't manufacture offences.
VeritasKit
by CcXXXXX0630
🔍 学术数据的测谎仪 — 10项统计鉴伪检测（Benford/GRIM/p-curve/Bootstrap），让造假无处遁形
the-board
by pamiray-m
Convene a panel of expert archetypes to pressure-test any decision and return a clear PROCEED / CONCERN / BLOCK verdict with reasons. Use when the user wants a board review, a second opinion, a go/no-go call, a pre-mortem, a red-team, or to vet/evaluate an idea, plan, strategy, launch, architecture, or risk. Triggers: "should I…", "is this a good idea", "review this decision", "get the board's take", "pros and cons", "what could go wrong", "vet this", "go or no-go". The board is composed of composite ADVISORY ARCHETYPES (e.g. The Strategist, The CFO, The Safety Reviewer) — it never impersonates real, named individuals.
tilde
by roktas
Use for Tilde deployment, provisioning, and home-management work. Trigger for Tilde prompt markers such as /tilde ... or $tilde ..., ~ ..., messages whose first word is tilde, or requests to create, initialize, deploy, update, diagnose managed state, adopt app/config files, customize data-layer policy, or work with Tilde home behavior.