Análise 5+2 de Ações — Método Fundamental
YichengYang-Ethan/meitou-5plus2-en
Análise fundamental 7-passos (5 objetivos + 2 subjetivos) para ações US
Ver no GitHubSkill para Claude Code que executa o método de análise fundamental 5+2 em qualquer ação americana. Classifica automaticamente o arquétipo da empresa e aplica o playbook correto (métricas financeiras, método de valuation, análise de moat). Retorna um briefing estruturado com avaliação de justo valor, tese de compra e riscos.
Instalação
npx convez-ai YichengYang-Ethan/meitou-5plus2-en1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Avaliação fundamentalista de ações americanas
- Comparação de valuation com múltiplos contextualizados
- Identificação de moat competitivo e qualidade de gestão
Veredito

Framework robusto e bem documentado que transpõe método consagrado (MeiTouJun) para Claude Code, com roteamento automático de arquétipos e adaptação de métricas. Prático para quem busca análise estruturada, sem ser recomendação de investimento; dados EOD do yfinance exigem validação antes de operações reais.
README
A tradução para o português deste README ainda está na fila.
meitou-5plus2 — the "5+2" stock-analysis method, as a Claude Code skill
A portable Claude Code skill that runs the 5+2 fundamental-analysis method on any US stock: give it a ticker → it pulls the hard numbers → works the 7 steps → outputs a research brief (full understanding + is it fairly valued + bull case vs bear case).
Method credit. The "5+2 method" was popularized by 美投君 (MeiTouJun), a Chinese-language YouTube channel on US equities. This repo is an independent distillation of the framework (steps, thresholds, judgement style). It contains none of the channel's videos or transcripts. Not investment advice.
What "5+2" is
7 steps, 5 objective + 2 subjective:
| Step | ||
|---|---|---|
| ① | Industry | objective |
| ② | Business model | objective |
| ③ | Management | objective |
| ④ | Financials | objective |
| ⑤ | Valuation | objective |
| ⑥ | Investment logic (why buy) | subjective |
| ⑦ | Investment risk (why not) | subjective |
The trick that makes it work on any ticker (not just well-known ones): the skill classifies
the company's archetype first (references/06-archetype-router.md) — 8 archetypes + a universal
fallback — then applies the right playbook (which financial metrics, which valuation method, which
moat lens). A bank gets P/B×ROTCE, a story stock gets SOTP-DCF, a loss-making consumption-SaaS gets
EV/Sales — automatically.
Quick start
# 1) clone
git clone <this-repo-url> meitou-5plus2
cd meitou-5plus2
# 2) install the one dependency (the data adapter uses yfinance only)
pip install -r requirements.txt # i.e. pip install yfinance
# 3) make it a Claude Code skill (symlink into your skills dir)
ln -s "$(pwd)" ~/.claude/skills/meitou-5plus2
# 4) (optional) sanity-check the data adapter on its own
python3 scripts/factsheet.py NVDA # human-readable
python3 scripts/factsheet.py NVDA --json # structured
Then in Claude Code:
/meitou NVDA
or just: "analyze NVDA with the 5+2 method."
What the data adapter gives you
scripts/factsheet.py (yfinance only, no other services) returns: valuation multiples (PE, PEG,
P/B, P/S, EV/Sales), WACC + reverse-DCF implied growth, analyst-target asymmetry, margins
(gross/operating/net/FCF), ROE/ROA/ROIC, LT Debt / Total Assets (the Buffett <40% line),
quarterly gross-margin trend, SBC / revenue (SaaS dilution), revenue/earnings growth,
RSI/MA/drawdown, and recent news. The qualitative parts (industry narrative, moat, management) are
left to the model + web search.
Layout
meitou-5plus2/
├── SKILL.md # the skill: 7-step framework + archetype routing + voice
├── requirements.txt # yfinance
├── scripts/factsheet.py # standalone data adapter (ticker -> factsheet)
└── references/
├── 06-archetype-router.md # 🧭 classify any company -> playbook (+ universal fallback)
├── 5plus2-rubric.md # the distilled rules for each of the 7 steps
├── 05-rating-system.md # rating / tier / target framework (+ coverage gate)
├── method.md # the method, in brief
└── voice.md # how to write the brief in the author's voice
Notes
- Needs Python 3.9+ and
yfinance. Data is delayed/EOD from public sources; double-check anything you'd trade on. - The skill is a framework, not a stock picker. It produces a structured read, not a guarantee.
Not investment advice. Reproduces an analytical framework only.