B-M-Capital-Research/honeclaw

최근 기록된 소스 활동: 2026년 10월 12일 04:19

SkillsMP 카탈로그 업데이트: 2026년 11월 5일 19:57

수집된 skills: 36

GitHub 스타: 751
GitHub 포크: 110

AI 어시스턴트로 설치

이 프롬프트를 복사해 사용 중인 AI 어시스턴트에 입력하세요.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/B-M-Capital-Research/honeclaw.

저장소 README

README 원본 파일: README.md

README 원본 보기

README 전체 펼치기

<img src="./resources/logo.svg" alt="HONE" width="160"> <h1 align="center">Hone · Your AI Infrastructure Research Assistant</h1> Start with a question. Understand the industry. Keep up with the companies you follow. Now with automated quantitative trading: <a href="https://github.com/B-M-Capital-Research/hone-quant">hone-quant</a>. Optical communications · Power · Memory & storage · Chips · Data centers English · <a href="./README_ZH.md">简体中文</a> <a href="https://hone-claw.com/chat">Start using Hone</a> · <a href="https://hone-claw.com">Website</a> · <a href="https://hone-claw.com/blog">Blog</a> · <a href="https://discord.gg/TyDNfYXDGF">Discord community</a> · <a href="#contact">Contact</a> --- ## New: hone-quant, automated quantitative trading built on honeclaw **Hone is more than an AI chat tool.** On top of honeclaw's capabilities, **[hone-quant](https://github.com/B-M-Capital-Research/hone-quant)** turns Hone's industry research into automated quantitative trading. Professional researchers maintain and operate Hone's industry ontology of the AI-infrastructure supply chain. AI agents rigorously generate strategies from it and execute them twice every US trading day. Every target weight, order and manual decision is recorded and auditable. <img src="./resources/readme/hone-quant-backtests.jpg" alt="hone-quant backtests on hone-claw.com/quant: the sector-first risk budget over three years and one year" width="100%"> | Sector-first risk budget, historical backtest | **3 years** (2023-10-05 → 2026-10-05) | **1 year** (2025-10-05 → 2026-10-05) | | --- | ---: | ---: | | Total return | **+535.88%** | **+70.69%** | | Annualised return | +85.24% | +71.01% | | Sharpe ratio | 2.10 | 1.59 | | Maximum drawdown | −28.01% | −24.57% | | Excess return vs QQQ | **+421.46%** | +45.68% | - **Researchers own the ontology.** It decides *what* is in scope: the sectors, the companies and how demand travels between them. hone-quant never picks stocks. - **AI agents turn it into strategy.** They set sector budgets first, then risk-based weights within each sector, and backtest every strategy on the same engine that trades. Each strategy is saved as an immutable version, and every plan explains each target. - **The engine keeps the discipline.** Review windows, quote and price-deviation checks, a 5% single-name cap, a turnover cap and no leverage. People can approve, cancel or pause at any time. Historical backtests on real market data (split- and dividend-adjusted daily prices, slippage, commissions and fees included), run on the same engine as the live paper account. The universe is today's ontology membership, so results carry survivorship bias and are likely optimistic. hone-quant trades a paper account only. Past and simulated performance do not guarantee future results. Not investment advice. **[Explore hone-quant →](https://github.com/B-M-Capital-Research/hone-quant)** --- **How is progress in AI changing the companies you follow?** New models, growing inference demand, and rising cloud capital expenditure flow through to chips, memory, optical interconnects, power, and data centers. Hone helps you connect those developments to individual companies: ask a question in plain language, explore its implications with industry research and professional financial data, and follow the companies that deserve ongoing attention. Hone has built an **ontology-based research foundation covering around 50 AI infrastructure companies**, spanning optical communications, power, memory and storage, chips, and related segments. This foundation supports industry context and company analysis in conversations, helping research follow relationships across the supply chain. <img src="./resources/readme/hone-investment-assistant.png" alt="Hone user-facing investment assistant: exploring how advances in AI models affect the supply chain and individual companies" width="100%"> The investment assistant: start with a natural-language question and keep exploring in the same conversation. Screenshot shown as a UI example; consult current sources when using the product. ## Turn AI headlines into research questions You can start without choosing a tool, writing a prompt template, or preparing a research outline. Open the investment assistant and ask what matters to you: - “How does growing inference demand affect optical modules, memory, and power?” - “How should I think about Micron's valuation, and which operating metrics should I track?” - “What does higher cloud capital expenditure mean for each company I follow?” - “Does this news change the company's long-term thesis, or reflect short-term sentiment?” Hone combines AI industry context, company research, and currently available data to help you examine transmission paths, key assumptions, and questions that need more evidence. Follow up on an answer, or use “Ask HONE” in a research panel to continue with that context. ## Follow companies and build an ongoing research routine Add companies to your watchlist or holdings to give your research a starting point. Interpret company news, earnings developments, and relevant events in the context of their businesses and industries. Review important updates in the app's notifications, then return to the assistant for further analysis. **Hone connects to professional financial databases** to support questions about financial statements, growth, profitability, cash flow, valuation, and market data. Move from “What happened?” to “What does it mean for the business?”, “What expectations does the valuation imply?”, and “Which data should I check next?” For a memory company, that might mean following pricing cycles, capacity, margins, and capital expenditure. For an optical communications company, it might mean examining product transitions, customer demand, and delivery schedules. Data coverage and refresh rates depend on the source. Feature availability also depends on account permissions and service configuration. ## Keep up with Serenity and other market voices on your phone Open **Research → Influencer Digest** (「研究 → 大V速报」) to browse public updates from supported authors such as **Serenity / 白毛**. The live digest currently **syncs every 15 minutes**, bringing scattered updates into one timeline so you can keep up with industry discussions throughout the day. - Read Chinese translations, expand English originals where available, and view images and publication times. - Open the author's original post and the translation or aggregation source to check the full context. - Expand “HONE Interpretation” to explore the distinction between opinions and facts, then continue your research with “Ask HONE.” <img src="./resources/readme/hone-serenity-mobile.jpg" alt="Hone mobile Influencer Digest showing Serenity updates, original images, source links, and Ask HONE" width="360"> Captured from the live user app in a mobile viewport. Author posts, original sources, and HONE interpretations are shown separately; an author's opinion does not represent Hone's judgment. ## Read the macro signals—and their impact on AI infrastructure The Macro Traffic Light helps you monitor the broader market and economic cycle, review changes in financial conditions and market risk, and inspect the reasoning behind the signals. Start with the signal, open the details, or ask a follow-up question. **Hone puts particular emphasis on how those changes affect AI infrastructure.** Interest rates, liquidity, and risk appetite matter through their effects on financing costs, cloud capital expenditure, data center construction, and upstream demand. Research also considers AI demand, commercialization, financing, and capital expenditure to examine the durability of industry growth. “What color is today's signal?” is a starting point. The next questions are which companies are affected, how the impact travels through the industry, whether business fundamentals have changed, and what further evidence is needed. ## Start using Hone 1. Open **[the Hone user app](https://hone-claw.com/chat)**, follow the sign-in prompts, and ask your first AI or company research question. 2. Add companies you follow or hold so future updates have a clear research focus. 3. Visit Research for the Macro Traffic Light and Influencer Digest, and Notifications for updates relevant to you. 4. When something deserves a closer look, select “Ask HONE” to continue in the conversation. Use Hone in your browser: explore in depth on a computer, or check updates and ask follow-up questions on your phone. ## Open source and self-hosting Honeclaw is the open-source project behind Hone. Built with **Rust + SolidJS / TypeScript**, it provides the application foundation for conversations, research, notifications, and access across channels. To explore the implementation, contribute, or self-host, start here: - [Complete wiki: installation, configuration, startup, and troubleshooting](./docs/wiki.md) - [Repository map and module entry points](./docs/repo-map.md) - [Development and collaboration guidelines](./AGENTS.md) - [Releases and installation packages](https://github.com/B-M-Capital-Research/honeclaw/releases) - [hone-quant: automated quantitative trading on the honeclaw ontology](https://github.com/B-M-Capital-Research/hone-quant) The professional financial databases, proprietary ontology research, and certain research assets used by the hosted product are not distributed with the public code. Self-hosting requires your own model and data service configuration; cloning the repository does not grant access to all hosted data or account features. Share feedback and contribute through [Issues](https://github.com/B-M-Capital-Research/honeclaw/issues), Pull Requests, or [Discord](https://discord.gg/TyDNfYXDGF). For more, visit [Bilibili: 巴芒投资](https://space.bilibili.com/224670487) and [YouTube: 巴芒投研美股频道](https://www.youtube.com/@%E5%B7%B4%E8%8A%92%E6%8A%95%E7%A0%94%E7%BE%8E%E8%82%A1%E9%A2%91%E9%81%93). To hone is to sharpen. We want Hone to help you sharpen your judgment through ongoing research, comparison, and review. ## Contact We are an investment research organization based in Singapore, and the team behind honeclaw and [hone-quant](https://github.com/B-M-Capital-Research/hone-quant). If these open-source projects interest you, write to us. You might want to use them, build on them or work with us. We also offer community consultation, investment research guidance and related services. - **E-mail:** [contact@honeclaw.app](mailto:contact@honeclaw.app) - **Website:** [hone-claw.com](https://hone-claw.com) This project is released under the [MIT License](./LICENSE). Hone's content is for research purposes and does not constitute investment advice.

저장소 README를 불러오는 중…

이 저장소의 skills

재무 및 투자 분석가 소프트웨어 개발자 시장조사 분석가·마케팅 전문가 회계사 및 감사원

직업 카테고리 9개 · 97% 분류됨

일반 사무원

管理当前用户的市场事件推送偏好,把中文自然语言映射到 notification_prefs 工具

원문 언어: 중국어

2026-08-28

특수 효과 아티스트 및 애니메이터

Generate dynamic portfolio screenshots, stock-analysis infographics, or general-style images. Real data must be fetched first

원문 언어: 영어

2026-03-22

pdf-understanding

재무 및 투자 분석가

Read and analyze user-uploaded PDFs, preferably using the system-extracted text fragments, and output summaries, key points, and risk notes

원문 언어: 영어

2026-03-22

수집된 skill 36개 중 36개를 표시합니다.