Analysis updated 2026-07-03 · repo last pushed 2026-06-30
Ask when Zheng Xi started favoring optical communication and get the exact interview quote plus his quarterly holdings.
Compare his tech-focused fund head-to-head with another manager's healthcare fund using his scoring framework.
Score a popular liquor index fund against his investment criteria to see if it matches his style.
Study his investment methodology through distilled principles each backed by his own words.
| lyra81604/zhengxi-views | forsy-ai/agent-apprenticeship | claudiodrews/memory-os | |
|---|---|---|---|
| Stars | 1,151 | 1,189 | 1,222 |
| Language | Python | Python | Python |
| Last pushed | 2026-06-30 | 2026-07-03 | 2026-06-10 |
| Maintenance | Active | Active | Active |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 3/5 | 4/5 |
| Audience | researcher | pm founder | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires an AI platform that supports agent skills (like Claude Code or Cursor) and some platforms need Python runtime to execute scripts and fetch live fund data.
This project, zhengxi-views, is an AI "skill" that lets you ask questions about a specific Chinese fund manager named Zheng Xi (from E Fund Management) and get answers backed by his actual public statements, not made-up text. Instead of an AI guessing what he might think about a topic, every answer traces back to a specific year, document, and original quote. It also scores any of China's roughly 27,000 public mutual funds against his investment framework, so you can see how closely a fund matches his style. The tool is built on three pillars: a corpus of Zheng Xi's published writings from 2012 to 2026 (quarterly reports, personal notes, media interviews), a distilled summary of his investment method where every principle is backed by his own words, and real fund data covering his 8 managed funds plus the broader market. When you ask a question, it searches the corpus for direct quotes. If the corpus doesn't cover a topic, it applies his documented methodology to reason through it, but always flags that output as inference, not his actual opinion. Python scripts handle corpus searching, on-demand fund data fetching, and the six-dimension scoring. This would appeal to investors, analysts, or students who follow Zheng Xi and want a reliable way to study his views and verify his track record. For example, you could ask "When did he start favoring optical communication?" and get the exact interview quote plus a check against his actual quarterly holdings. You could compare his tech-focused fund head-to-head with another manager's healthcare fund, or ask whether he'd buy a popular liquor index fund (the answer: almost certainly not, scoring 21/100 on his criteria). The project is designed as an Agent Skill that plugs into AI platforms like Claude Code, Tencent's WorkBuddy, Cursor, ChatGPT, and Gemini. Some platforms can run its Python scripts and fetch live data, others are limited to reading the pre-packaged corpus and fund snapshots. The maintainers are clear about boundaries: it's for research and learning only, not investment advice, and the cardinal rule is never fabricating statements he didn't make.
An AI skill that answers questions about Chinese fund manager Zheng Xi using his real public statements, and scores any of China's 27,000 mutual funds against his investment framework.
Mainly Python. The stack also includes Python.
Active — commit in last 30 days (last push 2026-06-30).
No license information is provided in the repository.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly researcher.
This repo across BitVibe Labs
Verify against the repo before relying on details.