# AlphaGBM/investment-masters: 15 Investor Methodologies as Markdown for AI Agents > The repository turns shareholder letters, memos and 13F filings into structured master profiles an AI agent can read. It is a knowledge base, not a screener, and its value depends on the files you feed your agent. **AlphaGBM/investment-masters** — Distill investment wisdom from Buffett, Dalio, Soros, Marks & more — 15 masters' methodologies + 13F tracking, built for AI agents. - Repository: https://github.com/AlphaGBM/investment-masters - Website: https://www.alphagbm.com/skills - Stars: 631 · Forks: 32 - Language: Unknown - License: MIT - Published: 2026-09-10 · Updated: 2026-09-10 · Language: en - Canonical page: https://hysenlabs.com/projects/alphagbm-investment-masters ## What Investment Masters Actually Packages The README frames the problem plainly: the useful public record in investing is scattered across shareholder letters, SEC filings, books, interviews and memos. Investment Masters collects that record for 15 named managers and rewrites it as structured profiles. Buffett, Dalio, Simons, Asness, Tepper, Soros, Ackman, Marks, Zhang Lei, Wood, Duan Yongping, Lynch, Druckenmiller, Liang Wenfeng and Linda Raschke each get a file under masters/. The audience is narrow and the repository says so. The subtitle names AI agents and human investors, and the first quick-start instruction targets Claude Code specifically. If you are looking for a screening tool or a portfolio tracker, this is the wrong shape of thing. What you get is prose organised so a language model can retrieve it: selection criteria, position sizing, risk control, exit conditions, and a citation trail the README claims exists for every assertion. The README contrasts this with what it calls AI roleplay tools, which it characterises as surface-level quotes drawn from LLM training data. Whether the profiles live up to that contrast depends on the individual files, which the README does not reproduce. ## How the Skill Files and 13F Table Fit Together The architecture is deliberately thin, and that is the point. There is no server, no database and no build step. The repository root holds README.md, SKILL.md, LICENSE and a masters/ directory of Markdown profiles. SKILL.md is the agent-facing definition; the profiles are the payload an agent reads when a query matches. Data flow is retrieval, not computation. You ask a question, the agent loads the relevant profile file, and answers from that text. The README's example table maps queries to behaviour: comparing Dalio and Marks on risk produces a side-by-side analysis of their risk philosophies, and asking about Ackman's latest 13F changes sends the agent to EDGAR for the Pershing Square filing. That second example is where the design gets interesting. The repository does not mirror filing data. It stores CIK numbers and links to SEC EDGAR, which is free and public. Bridgewater is listed as 0001350694, Berkshire as 0001067983, Appaloosa as 0001656456, Pershing Square as 0001336528, Soros Fund as 0001029160, Hillhouse as 0001510057 and ARK Invest as 0001803918. Freshness therefore depends on the agent fetching EDGAR at query time, not on repository updates. The README also notes that Renaissance Medallion and AQR's internal funds are proprietary and absent from 13F, so two of the fifteen masters have no holdings data at all. ## Installing It as a Claude Code or Cursor Skill There is no package manager step. Installation is a git clone into the directory your editor or agent reads skills from. For Claude Code, the README gives this command, which places the repository under .claude/skills/investment-masters in the current project: ```bash git clone https://github.com/AlphaGBM/investment-masters.git .claude/skills/investment-masters ``` For Cursor, the same clone targets a different path: ```bash git clone https://github.com/AlphaGBM/investment-masters.git .cursor/skills/investment-masters ``` After cloning, the README says to ask the AI directly. These are the prompts it lists, and the expected result is a structured profile covering principles, sizing, risk control, the latest 13F and takeaways: ```text Distill Buffett's investment methodology Compare Dalio and Marks on risk management What did Bridgewater buy last quarter? ``` The first prompt should return content drawn from masters/buffett.md. The third depends on network access to EDGAR. If your agent runs sandboxed without outbound requests, that query will fail while the other two still work, because only the 13F questions need the live filing. The README does not document a fallback for that case. ## Where the Repository Stops Being Useful The 13F table is the weakest part, and the README is honest about why. Filings are quarterly, so any holdings answer is up to three months stale by construction, and 13F forms disclose long positions without short positions or most derivatives. A reader asking what a fund is doing right now will get a lagged snapshot and no visibility into hedges, which matters for a manager like Ackman whose 2020 CDS trade the README itself cites as an example of event-driven positioning. Coverage is also uneven. Seven managers have CIK entries. The other eight do not, and two of those, Renaissance and AQR, are excluded for structural reasons rather than oversight. The README states that Renaissance Medallion and AQR's internal funds are proprietary and not in 13F. There is a second limitation the README does not address. Methodology extraction produces text, not rules. The five common principles it lists, systems over intuition, risk management over stock picking, clear thesis with willingness to be wrong, cycle awareness, and long-term over short-term, are descriptions of how managers think. Nothing in the repository converts them into position sizes, entry thresholds or exit triggers you could execute. Treating the profiles as a system rather than as reading material is a misuse the project cannot prevent. ## Against a Quant Factor Library The natural alternative is a factor research library such as the published AQR paper set, or an open quant framework where signals are expressed as code. Those give you something Investment Masters does not: parameters you can run. A momentum factor definition can be computed over a price series and compared against a benchmark. A profile of Asness's methodology cannot. The difference in approach is the unit of knowledge. Quant libraries encode decisions as functions with inputs and outputs. Investment Masters encodes them as prose with citations, and it is explicit that the same managers appear in both worlds. AQR is one of the fifteen profiles and also the source of the 200+ research papers the README attributes to it. If you want the factor, read the papers. If you want the reasoning behind the factor, in a form an agent can quote, the profile is the shorter path. The trade-off is verifiability. A backtest tells you whether a rule worked on historical data. A methodology profile tells you what a manager said, and the README's claim that every assertion carries a source citation is the only quality control available. There is no test suite in the repository layout, and no CI configuration is listed among the top-level entries. ## Licence, Maintenance and Upgrade Cost The licence is MIT, per the LICENSE file and the badge in the README. That permits commercial use, modification and redistribution provided the copyright notice and permission notice are retained. It says nothing about the underlying source material. The profiles summarise copyrighted books, shareholder letters and memos, and MIT covers the repository's own text, not the works it discusses. If you plan to republish profile content, that distinction is worth raising with a lawyer rather than assuming the repository licence settles it. The last push to the default branch was on 2026-05-27. The repository is not archived. Upgrading is a git pull, since there is no dependency manifest, no lockfile and no release artefact. The cost of an upgrade is the cost of re-reading diffs in Markdown files you may have edited locally, which is a real friction if you tuned a profile for your own prompts. The README does not document a rollback path, and with no tagged releases retrieved, there is no version to pin to. ## Conclusion Adopt it if you already run Claude Code or Cursor and want sourced, citable methodology text in the context window instead of model-generated Buffett impressions. Skip it if you need live prices, position sizing maths or a backtest engine, because the repository contains none of those. Before relying on it, open masters/buffett.md and masters/howard_marks.md and check whether the citations are specific enough for your use, then confirm each CIK against the SEC EDGAR page rather than the README table. ## FAQ ### What is AlphaGBM/investment-masters? It is an MIT-licensed repository that distills 15 fund managers' methodologies into structured Markdown profiles, with a 13F tracking table linking to SEC EDGAR filings. The README describes it as built for AI agents and human investors. ### How do I install investment-masters for Claude Code or Cursor? Clone the repository into the skills directory your tool reads. The README gives git clone https://github.com/AlphaGBM/investment-masters.git .claude/skills/investment-masters for Claude Code and the same command with .cursor/skills/investment-masters for Cursor. ### Does investment-masters give live portfolio holdings? No. It stores CIK numbers and links to SEC EDGAR, and the agent fetches filings at query time. Because 13F filings are quarterly, any holdings answer reflects the most recent filing rather than current positions, and Renaissance Medallion and AQR's internal funds are not in 13F at all. ### Which managers are covered by investment-masters? The README lists fifteen: Bridgewater (Dalio), Buffett, Renaissance (Simons), AQR (Asness), Tepper, Soros, Ackman, Howard Marks, Hillhouse (Zhang Lei), ARK (Wood), Duan Yongping, Peter Lynch, Druckenmiller, Liang Wenfeng and Linda Raschke. Seven of them have CIK entries in the 13F table. ## Sources - [AlphaGBM/investment-masters on GitHub](https://github.com/AlphaGBM/investment-masters) - [Issues](https://github.com/AlphaGBM/investment-masters/issues) - [License: MIT](https://github.com/AlphaGBM/investment-masters/blob/main/LICENSE) - [Project website](https://www.alphagbm.com/skills) - [README](https://github.com/AlphaGBM/investment-masters/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/alphagbm-investment-masters