value-investor-skill
English · 简体中文 · 日本語 · 한국어 · Deutsch · Français · Español
From a stock name to a checkable valuation
A free local research desk: fetch public filings, inspect cash generation, test three valuation methods, and save what would change your mind. One AI skill, eight research workflows. Python 3.9+, no model API or pip dependencies for the desk.
git clone https://github.com/AhaSignals/value-investor-skill.git
cd value-investor-skill
python3 scripts/investor.py desk
Open the printed local URL. Choose a frozen Microsoft, Coca-Cola or Caterpillar example offline, or enter a supported US company for fresh sources. Load an explicitly illustrative scenario to learn the controls, then review and change its assumptions. Your assistant can handle the same workflow from a company name when it has the full project, Python and network access.
3 real financial datasets · 3 valuation methods · reverse DCF · sensitivity chart/table · thesis snapshots · optional PIT checks. Normalized cash and valuation assumptions need review; the desk does not invent a fair value from a ticker. AI service fees, if used, remain separate. Walk through the three companies · Models and limits · AhaSignals research desk guide.
Just enter a US stock name or ticker
Research MSFT. Fetch public filings and historical prices, explain profits versus cash, and show the strongest counterevidence.
With the full skill installed in a host with Python and network access, this starts the source workflow automatically. No JSON intake or API key. In an ordinary chat, source access depends on the app; pasted instructions do not install the downloader.
# From the cloned or extracted project directory:
python3 scripts/investor.py research Microsoft --out output/msft
Open output/msft/research.html. The report links financial statements, descriptive historical P/E and daily price context; your assistant reads it and the filings for deep business research. US nonfinancial companies first. Missing values remain unknown. Coverage and replay.
Try a real case: Microsoft: profit rose, cash after capex fell. A frozen five-year SEC extract reproduces the financial screen offline. A separate live run captured 1,260 daily bars and 61 monthly P/E samples; availability varies. Inspect the public report.
Is this stock worth its price?
8 value-investing workflows in 1 AI skill: scan financial statements, compare historical P/E, inspect price trends, understand the business, test valuations, check price assumptions, challenge your thesis, and track evidence changes.
For Claude, ChatGPT (GPT), DeepSeek, Grok, Gemini and Kimi. Start with a company name or a question. Copy the instructions into chat with no installation, or install the full skill for local Python calculations and saved research.
3 real SEC financial cases · 6 chat assistants · 7 language guides · 3 fictional case studies · 4 demo illustrations · HTML, Markdown and JSON reports.
Free, MIT, no project signup. Your AI service's account, tool access and fees still apply.
Try the 30-second browser demo · Choose your assistant · 中文 · By AhaSignals
Start with the assistant you already use
| You use | Start here |
|---|
| Claude, ChatGPT, DeepSeek, Grok, Gemini or Kimi chat | Copy these ready-to-paste instructions, then ask about a company. No terminal needed. |
| Claude with Skills enabled | Upload the skill ZIP. Steps. |
| Claude Code | Run the two plugin commands below. |
| Gemini CLI or Kimi Code | Install the same skill with --host gemini or --host kimi. Steps. |
| GPT in Codex | Ask it to install this repository as a skill. Steps. |
| DeepSeek in a skill-capable app | Use the Claude Code or Codex route for your configured host. Steps. |
In Claude Code:
/plugin marketplace add AhaSignals/value-investor-skill
/plugin install value-investor@value-investor-skill
Then ask:
Show me the fictional $40 example. Change one valuation assumption, then withdraw its supporting evidence. Explain what changes and why.
For your own research:
Research [company]. How does it make money, what does its price assume, and what evidence would change the answer?
The assistant supplies reasoning and source access. Python calculates supplied scenarios and preserves revisions when execution and writable files are available. Ordinary chat uses the same research method without promising a persistent ledger. Capabilities and troubleshooting.
Python 3.9+; no pip packages, API key or model calls for the fictional demo.
git clone https://github.com/AhaSignals/value-investor-skill.git
cd value-investor-skill
python3 scripts/doctor.py
python3 scripts/investor.py demo --out output/first-demo
This is the beginning of the README. Read the rest at the original. Read the full README ↗
