InvestSkill is a collection of that turn any AI assistant into an institutional-quality investment analyst.
Overview
InvestSkill is a collection of that turn any AI assistant into an institutional-quality investment analyst.
README
What is InvestSkill?
InvestSkill is a collection of 34 structured analysis frameworks that turn any AI assistant into an institutional-quality investment analyst. There is no runtime — every skill is a prompt that works in Claude Code, Cursor, Gemini CLI, GitHub Copilot, ChatGPT, or any other LLM.
Nothing to sign up for, nothing to pay for. No API key, no data-vendor subscription, no billing setup — you bring the AI assistant you already use (a free tier or a local model works too) and InvestSkill is just markdown. See No API keys, no cost.
Ask your AI: "Evaluate AAPL using the stock-eval framework"
Get back: Piotroski score · ROIC · moat rating · buy/hold/sell signal
A fifteen-lesson field guide, in plain English and Traditional Chinese: Part I teaches the concepts behind every skill — reading a balance sheet, valuing a business, holding a portfolio — and Part II the practical foundations: accounts and orders, an ETF core, taxes (including for non-US investors), earnings season, the psychology that protects the plan, the macro and Fed cycle, and options for stock holders. No finance degree required.
Start learning →
Quick Start
One command, any AI agent — pick yours on the website:
# Claude Code
curl -fsSL https://raw.githubusercontent.com/yennanliu/InvestSkill/main/install.sh | bash -s -- -a claude
# Swap the agent: cursor · copilot · gemini · codex · opencode · any
curl -fsSL https://raw.githubusercontent.com/yennanliu/InvestSkill/main/install.sh | bash -s -- -a cursor
The installer copies every framework into .investskill/prompts/ and wires up your agent’s own entry point — nothing else runs on your machine:
-a |
Agent | Wires up |
|---|---|---|
claude |
Claude Code | .claude/skills//SKILL.md (slash commands) |
cursor |
Cursor | .cursor/rules/investskill.mdc |
copilot |
GitHub Copilot | .github/copilot-instructions.md |
gemini |
Gemini CLI | GEMINI.md |
codex |
Codex | AGENTS.md |
opencode |
OpenCode | AGENTS.md |
any |
ChatGPT, Claude.ai, Ollama, … | .investskill/prompts/ only — paste and go |
Add -g to install for your user instead of one project, -d DIR to target another directory, or -h for all options. Existing instruction files are appended to, never overwritten, and re-running is a no-op. Read it first if you’d rather not pipe to bash: install.sh.
Claude Code via the plugin marketplace — full slash-command support:
claude
/plugin marketplace add yennanliu/InvestSkill
/plugin install us-stock-analysis@invest-skill
/us-stock-analysis:stock-eval AAPL
Or clone the repo — every framework in prompts/, use it anywhere:
git clone https://github.com/yennanliu/InvestSkill.git
cd InvestSkill
# Cursor
cursor .
# → @prompts/stock-eval.md Evaluate Apple
# Gemini CLI
gemini
# → @prompts/stock-eval.md Evaluate Apple
# Any LLM — paste the prompt file directly
cat prompts/stock-eval.md | pbcopy
Full platform guides: Claude Code · Cursor · Gemini CLI · Ollama
No API Keys, No Cost
InvestSkill is completely free to use — there is nothing to buy, register, or configure.
| API keys | None. No OpenAI / Anthropic / Alpha Vantage / Polygon key. The plugin never calls an endpoint — there is no code to call one with. |
| Subscriptions | None. No paid tier, no seats, no usage metering, no account. |
| Market-data fees | None. No Bloomberg, no paid data feed. The AI reads public filings and free sources — or you paste the numbers in yourself. |
| Install / runtime | None. Every skill is a plain markdown prompt. Nothing runs on your machine, so nothing phones home. |
| License | MIT. Free to use, fork, and modify — for personal or commercial work. |
The only thing you need is an AI assistant you already have. That can be a paid plan (Claude Code, Cursor, Copilot), a free tier (ChatGPT, Gemini, Claude.ai), or a fully offline local model via Ollama — in which case the total cost of running every framework is literally zero.
Because there’s no API, there’s also no telemetry: InvestSkill never sees your tickers, your holdings, or your analysis. See Data & Accuracy for where the numbers actually come from.
In Traditional Chinese, too. Add --lang zh-TW to any framework and the whole analysis comes back in 繁體中文 — headings, narrative and tables — while the Data & Sources labels, the signal box and its JSON footer stay in English so tools can still read them. The site has a Chinese Skill Reference and a Non-US Investor Guide.
Bring your own data. The filing skills (10k-digest, financial-report-analyst, fact-check) carry a keyless SEC EDGAR recipe a tool-enabled assistant can follow by itself. If yours can’t browse, two optional zero-dependency scripts fetch the primary source for you to paste — node scripts/fetch-edgar.js AAPL --form 10-K saves the filing as text, node scripts/fetch-fundamentals.js AAPL builds a reconciled statement pack from the SEC’s XBRL facts. Neither is part of the plugin. Details on the Data & Accuracy page.
The 34 Frameworks
Core Stock Analysis
| Skill | What it produces |
|---|---|
stock-eval |
Piotroski F-Score, ROIC, quality rating, go/no-go signal |
technical-analysis |
MA chart (30/60/90/200/365d) with trade entry/target/stop, chart patterns, RSI/MACD, MTF alignment, Ichimoku |
stock-valuation |
P/E · P/S · EV/EBITDA · comparable company multiples |
economics-analysis |
Macro indicators, recession probability, rate sensitivity |
compare |
Compare 2–5 stocks head to head on one table — business quality, growth, valuation, balance sheet, momentum, risks and catalysts — then argue the pick: which one, why, what would flip it, and whether owning more than one adds anything |
Financial Reports
| Skill | What it produces |
|---|---|
financial-report-analyst |
10-K / 10-Q key findings, red flags, accounting quality |
10k-digest |
Structured markdown digest — abstract, section summaries, metrics table, refs (EN / 繁中) |
earnings-call-analysis |
Management tone, guidance delta, hidden risks |
forensic-accounting |
Earnings-quality forensics — Beneish M-Score, Altman Z, Sloan accruals ratio, cash-vs-earnings divergence, SBC-adjusted FCF, revenue-recognition and capitalisation checks, auditor and related-party flags, scored as a Manipulation Risk Score 0–10 |
credit-check |
Balance-sheet solvency and refinancing risk — debt maturity wall, interest coverage trend, leverage vs. covenants, liquidity runway, ratings and spreads, the cost of refinancing at today’s rates, and an equity-holder verdict, scored as a Credit Health Score 0–10 |
Market Monitoring
| Skill | What it produces |
|---|---|
insider-trading |
SEC Form 4 patterns, net buy/sell sentiment |
institutional-ownership |
13F holdings changes, smart money flows |
dividend-analysis |
Payout safety score, yield trap detection |
short-interest |
Short ratio, days-to-cover, squeeze probability |
Advanced Research
| Skill | What it produces |
|---|---|
competitor-analysis |
Moat score, Porter’s Five Forces, market share |
industry-map |
Supply/value-chain graph (upstream→downstream), chokepoints, margin-pool migration, second-order ideas |
options-analysis |
Greeks, IV rank, earnings play strategy selection |
portfolio-review |
Allocation health, concentration risk scoring, correlation analysis, tax-loss harvesting, rebalancing plan |
sector-analysis |
Sector rotation signals, valuation tables, seasonality calendar, momentum scoring |
stock-screener |
Multi-ticker ranking across valuation, quality, momentum, sentiment, and growth; leaderboard + top picks + avoid list |
catalyst-calendar |
Forward-looking 90-day event calendar: earnings, macro events, catalysts, impact scoring, event-driven strategies |
bear-case |
Deliberate short-seller red-team: bear thesis, accounting red flags, downside target, thesis-killers (counterevidence for any bull thesis) |
position-ladder |
Staged entry ladder + trim/re-add cycle for a single holding: share-count floor/ceiling, blended cost basis, wash-sale flags, total-return-vs-buy-and-hold check, thesis-break gate |
thesis-tracker |
Write, save, and re-check an investment thesis — KPIs with thresholds, invalidation triggers, catalysts, a pre-mortem, and a decision log; --update re-reads the saved file against new data and returns INTACT / WEAKENED / BROKEN |
etf-analysis |
ETF / index-fund due diligence — expense ratio vs. category, tracking difference, liquidity, holdings concentration and tilt, overlap with your other positions, distribution and capital-gains history, structure warnings, and an ETF-vs-top-holdings comparison, scored as an ETF Fitness Score 0–10 |
earnings-preview |
The before-earnings skill — consensus vs. whisper, 8-quarter beat rate and post-print move distribution, options-implied vs. realized move, what the current price already assumes, the KPIs to watch, and a three-scenario grid (beat-and-raise / beat-and-lower / miss) with expected reaction and a position rule for each |
tax-lens |
US tax mechanics for a position or portfolio — short- vs. long-term treatment, wash-sale window check, qualified-dividend holding-period test, lot selection (specific-ID vs. FIFO), tax-loss-harvesting pairs, account placement, estimated annual tax drag — plus a --non-us module (W-8BEN, dividend withholding and treaty rates, capital-gains treatment, US estate-tax exposure, UCITS alternatives). Educational only, never tax advice |
risk-stress-test |
Portfolio and position risk report — beta-weighted exposure, historical scenario replay (2008, March 2020, 2022 rate shock, 2025 tariff shock), parametric VaR / CVaR at 95 / 99 %, max-drawdown estimate, correlation-spike scenario, rate / USD / oil sensitivity, liquidity (days to exit at 20 % of ADV), and a Risk Budget Score 0–10 |
trade-postmortem |
Review a closed trade or a year of trades — process vs. outcome matrix, plan adherence, the biases that showed up, what the pre-trade skills said, and the one lesson for the journal; reads closed thesis-tracker files |
Meta & Output
| Skill | What it produces |
|---|---|
full-report |
Runs 5 / 10 / every applicable framework (--depth quick / standard / comprehensive, --skip to leave modules out) and saves a standalone HTML report |
report-generator |
Converts any analysis into a professional HTML/PDF report |
chart-master |
Mermaid · ASCII · Chart.js visualizations from financial data |
result-validator |
Scores any analysis on data quality, methodology, and signal consistency |
learning-coach |
Explains any InvestSkill output like a mentor — every metric in plain words, why it matters, its good / bad range, and the lesson that teaches it — then asks 3–5 Socratic questions and “what would change your mind?”. --level beginner / intermediate, --lang zh-TW, and a --quiz mode that drills a Learning lesson |
fact-check |
Claim-level verification of any report or data set — extract every figure and factual claim, check each against a primary source (SEC filing, company IR, FRED, exchange data, or the user’s own document), recompute derived numbers, mark ✅ verified / ⚠️ mismatch / ❓ unverifiable, and re-issue the report with inline citations and a References section; Verification Score 0–10 |
Aliases (redirects)
Three earlier skills were absorbed into larger ones. They are still installed so old references keep working, but they are aliases, not frameworks, and are not counted above.
| Alias | Redirects to |
|---|---|
fundamental-analysis |
stock-eval (statement-level deep dive is a section of it) |
dcf-valuation |
stock-valuation (DCF is Method 1 of its multi-method model) |
research-bundle |
full-report (use --depth quick / standard / comprehensive) |
Example Workflows
5-minute stock screen
/stock-eval NVDA
/stock-eval AMD
→ Piotroski scores + quality rating for quick go/no-go
Complete due diligence
/stock-eval AAPL
/stock-valuation AAPL --methods all
/competitor-analysis AAPL
/financial-report-analyst AAPL 10-K
→ Full investment thesis in one session
Earnings season playbook
/stock-eval TICKER ← pre-earnings baseline
/earnings-call-analysis TICKER ← post-earnings [paste transcript]
/options-analysis TICKER --earnings ← vol expectations + strategy
→ Complete earnings thesis
Export a professional report
/full-report AAPL
→ Saves output/AAPL_report_2026-05-12.html
(hero header · metric cards · interactive Chart.js · signal block)
Output Format
Every skill ends with a standardized Investment Signal Block:
╔══════════════════════════════════════════════╗
║ INVESTMENT SIGNAL ║
╠══════════════════════════════════════════════╣
║ Signal: BULLISH / NEUTRAL / BEARISH ║
║ Confidence: HIGH / MEDIUM / LOW ║
║ Horizon: SHORT / MEDIUM / LONG-TERM ║
║ Score: X.X / 10 ║
╠══════════════════════════════════════════════╣
║ Action: BUY / HOLD / SELL ║
║ Conviction: STRONG / MODERATE / WEAK ║
╚══════════════════════════════════════════════╝
HTML reports render this as a styled dark panel with a score progress bar, ghost watermark text, and a teal/sky gradient accent — see report-generator for the full design system.
Platform Support
| Platform | Setup | How it works |
|---|---|---|
Every platform installs with the same one-command installer — only the -a agent changes (see Quick Start). |
| Platform | Setup | How it works |
|---|---|---|
| Claude Code | install.sh -a claude, or the plugin marketplace |
38 native slash commands (34 frameworks + 3 aliases + 1 output tool) |
| Cursor IDE | install.sh -a cursor |
Auto-loads .cursor/rules/investskill.mdc |
| Gemini CLI | install.sh -a gemini |
Auto-loads GEMINI.md |
| GitHub Copilot | install.sh -a copilot |
Auto-loads .github/copilot-instructions.md |
| Codex / OpenCode | install.sh -a codex / -a opencode |
Auto-loads AGENTS.md |
| Ollama (local models) | install.sh -a any, run a local model |
Offline open-source models, no API key — guide |
| ChatGPT / Claude.ai / any LLM | install.sh -a any, paste a prompts/*.md file |
Platform-agnostic markdown |
Learn the Frameworks
New to investing, or unsure which skill to reach for? Start here:
| Guide | What it gives you |
|---|---|
| Concepts | The mental models behind the metrics — quality vs. value vs. growth, intrinsic value, moats, risk, and the signal block |
| Glossary | Plain-English definitions, formulas, and good/bad ranges for every metric the skills emit |
| Choose a Skill | Map your goal to the right framework, plus comparisons of the ones that overlap |
| Use Cases | End-to-end journeys by investor type — income, first-timer, earnings, macro, skeptic |
| Data & Accuracy | Where the numbers come from, how to spot hallucinations, and how to validate output |
Documentation
| Resource | Description |
|---|---|
| Live Docs Site | Full documentation — light / dark theme, English and 繁體中文 |
| Cookbook | Walkthrough examples and use cases |
| Skill Reference | All 34 frameworks, one browsable page each |
| Claude Code Guide | Plugin install + all slash commands |
| Cursor Guide | Auto-loading rules + @prompts/ usage |
| Gemini CLI Guide | File references + multi-framework chains |
| Ollama Guide | Local open-source models, offline & no API key |
| Adding Skills | Contributor walkthrough — new-skill.js scaffolds and wires a skill |
| FAQ | 45+ answers on install, platforms, usage and versions |
| Changelog | Version history |
Contributing
See ADDING-NEW-SKILLS.md for the full process. The short version:
node scripts/new-skill.js --category --title "…" --desc "…"— scaffoldsSKILL.mdwith the full output contract and wires the skill into the site, READMEs, cross-AI configs, and CHANGELOG- Write the analysis in
plugins/us-stock-analysis/skills//SKILL.md, thennode scripts/sync-prompts.js—prompts/.mdis generated from it (skills are auto-discovered; noplugin.jsonchange) - Run
npm test— structure, prompt sync, skill contract, counts, and install script must all pass
Open an issue to report bugs · Start a discussion to propose features.
Version: latest — see CHANGELOG · Frameworks: 34 analysis frameworks (+ 3 aliases, 1 output tool) · Skills: 38 · Platforms: 7 · License: MIT · Tests: all passing
Recommended Tools
Try a different keyword or remove a filter.
Install
npx skillfish add yennanliu/investskill