name

like-rich

description

Analyze, screen, value, compare, and report on investment candidates using the public mental models of great investors including Warren Buffett, Charlie Munger, Bill Ackman, George Soros, Jim Rogers, Benjamin Graham, Philip Fisher, John Paulson, Jesse Livermore, and Jim Simons. Use when asked in English or Korean to find stocks, build a watchlist, calculate intrinsic value, assess catalysts, evaluate macro reflexivity, detect value traps, compare investment styles, or write an evidence-based investment research report.

Like Rich Skill

Use this skill to analyze public equities and other liquid investment candidates through a multi-investor mental model stack. The skill combines quality, value, growth, special situation, macro, trend, trading-risk, and quantitative anomaly lenses.

This skill is for research support only. Do not present outputs as personalized investment advice, trading signals, guarantees, or claims that any named investor would buy or sell a security. Use "mental model inspired by..." rather than impersonating or attributing unsupported views.

Core Workflow

Clarify scope only when needed. Otherwise use reasonable defaults.
- Default language: English.
- Korean support: respond in Korean when the user writes in Korean or explicitly asks for Korean output. Keep Korean user-facing documentation in separate docs/*_KO.md pages, not side-by-side inside English pages.
- Default markets: KRX, NYSE, Nasdaq, NYSE American, major ADRs, and liquid global large caps.
- Default horizon: 3 to 10+ years unless the user asks for event-driven or trading analysis.
- Default output: ranked candidate table plus short company-by-company notes.
Gather current and primary-source data.
- Use current price, market cap, shares, financial statements, filings, IR materials, and material news.
- For time-sensitive facts, browse current sources and cite access dates.
- Prefer filings, annual reports, exchange/regulator pages, company IR, audited statements, and reputable financial data.
Choose the relevant mental model stack.
- For long-term quality: Buffett, Munger, Fisher, Ackman.
- For cheap downside protection: Graham, Buffett.
- For concentrated compounders or activist catalysts: Ackman, Fisher, Munger.
- For macro, commodities, currencies, or reflexive bubbles: Soros, Rogers.
- For crisis, distressed, or event-driven asymmetry: Paulson, Graham, Soros.
- For price action, risk control, and crowd behavior: Livermore.
- For factor evidence, base rates, and anomaly checks: Simons.
- Read references/great-investor-mental-models.md before any Top-N screen, full report, intrinsic-value analysis, or style-specific judgment.
- Use references/investor-mental-models.md only as a quick routing index.
- For Korean user-facing explanations of the framework, read docs/README_KO.md and docs/GREAT_INVESTOR_MENTAL_MODELS_KO.md as separate localized pages.
Score each candidate with the seven-axis framework.
- Business quality and moat.
- Management, incentives, and capital allocation.
- Financial strength and cash conversion.
- Growth runway and reinvestment.
- Valuation, downside protection, and margin of safety.
- Catalyst, reflexivity, and market psychology.
- Evidence quality, base rates, and quant sanity checks.
- Read references/evaluation-framework.md for scoring details.
Value candidates when the user asks for a Top-N list, ranking, intrinsic value, fair value, buy zone, margin of safety, or a full report.
- Read references/valuation-playbooks.md.
- Include current price, base-case intrinsic value per share, buy-below price, upside/downside, valuation method, confidence, and source dates.
- Use DCF/owner earnings for durable businesses, asset value for deep value, sum-of-the-parts for conglomerates, event-weighted value for special situations, and normalized cycle value for cyclicals.
- If a valuation is unreliable, mark it "inconclusive" and explain what evidence is missing.
Classify the opportunity.
- Compounder: high-quality business that can compound intrinsic value.
- Deep value: statistically cheap with downside protection.
- Special situation: catalyst-driven or event-driven asymmetry.
- Macro/reflexive: driven by liquidity, policy, commodity cycle, currency, or narrative feedback loops.
- Quant candidate: survives factor, base-rate, and anomaly checks.
- Watchlist: good but no margin of safety.
- Reject: unclear economics, weak evidence, poor risk/reward, or unacceptable governance.
Write the report.
- Include executive summary, methodology, ranking table, investor-model matrix, valuation table, company sections, risks, falsification triggers, and sources.
- For each serious candidate, include primary, secondary, and opposing investor lenses, the calculations used by those lenses, and the facts that support or weaken each lens.
- Avoid hype. Separate facts, assumptions, and judgments.
- State uncertainty clearly and include what would change the conclusion.
Verify quality before finalizing.
- Check ticker, exchange, currency, fiscal period, share count, one-off items, leverage, source dates, and whether current price data is fresh.
- Do not fabricate citations or fill missing data with invented numbers.

Required Output for Ranked Lists

Every Top-N or ranked candidate answer should include:

Column	Requirement
Rank	Ordered by risk-adjusted opportunity, not only quality.
Ticker/company	Include exchange and currency when useful.
Current price	Include source and access date.
Intrinsic value or fair value	Include method and confidence.
Buy-below price	Apply explicit margin of safety.
Upside/downside	Use current price versus intrinsic/fair value.
Best-fit investor lenses	Name the strongest mental models.
Composite score	Use the seven-axis framework.
Thesis	One concise, evidence-backed reason to care.
Main risk	One concise reason the thesis could fail.

Supporting Files

Investor models: references/investor-mental-models.md
Detailed investor mental models: references/great-investor-mental-models.md
Evaluation framework: references/evaluation-framework.md
Valuation methods: references/valuation-playbooks.md
Data sourcing: references/data-source-guide.md
Report writing: references/report-writing-guide.md
Red flags: references/risk-red-flags.md
Report template: templates/research_report_template.md
Company card schema: templates/company_card_schema.json
Korean user guide: docs/README_KO.md
Installation guide: docs/INSTALL.md
Korean installation guide: docs/INSTALL_KO.md
Korean mental-model guide: docs/GREAT_INVESTOR_MENTAL_MODELS_KO.md
Candidate scoring script: scripts/score_candidates.py
Report builder script: scripts/build_report_from_cards.py
Source validator: scripts/validate_sources.py

Like Rich Skill

Core Workflow

Clarify scope only when needed. Otherwise use reasonable defaults.

Default language: English.
Korean support: respond in Korean when the user writes in Korean or explicitly asks for Korean output. Keep Korean user-facing documentation in separate docs/*_KO.md pages, not side-by-side inside English pages.
Default markets: KRX, NYSE, Nasdaq, NYSE American, major ADRs, and liquid global large caps.
Default horizon: 3 to 10+ years unless the user asks for event-driven or trading analysis.
Default output: ranked candidate table plus short company-by-company notes.

Gather current and primary-source data.

Use current price, market cap, shares, financial statements, filings, IR materials, and material news.
For time-sensitive facts, browse current sources and cite access dates.
Prefer filings, annual reports, exchange/regulator pages, company IR, audited statements, and reputable financial data.

Choose the relevant mental model stack.

For long-term quality: Buffett, Munger, Fisher, Ackman.
For cheap downside protection: Graham, Buffett.
For concentrated compounders or activist catalysts: Ackman, Fisher, Munger.
For macro, commodities, currencies, or reflexive bubbles: Soros, Rogers.
For crisis, distressed, or event-driven asymmetry: Paulson, Graham, Soros.
For price action, risk control, and crowd behavior: Livermore.
For factor evidence, base rates, and anomaly checks: Simons.
Read references/great-investor-mental-models.md before any Top-N screen, full report, intrinsic-value analysis, or style-specific judgment.
Use references/investor-mental-models.md only as a quick routing index.
For Korean user-facing explanations of the framework, read docs/README_KO.md and docs/GREAT_INVESTOR_MENTAL_MODELS_KO.md as separate localized pages.

Score each candidate with the seven-axis framework.

Business quality and moat.
Management, incentives, and capital allocation.
Financial strength and cash conversion.
Growth runway and reinvestment.
Valuation, downside protection, and margin of safety.
Catalyst, reflexivity, and market psychology.
Evidence quality, base rates, and quant sanity checks.
Read references/evaluation-framework.md for scoring details.

Value candidates when the user asks for a Top-N list, ranking, intrinsic value, fair value, buy zone, margin of safety, or a full report.

Read references/valuation-playbooks.md.
Include current price, base-case intrinsic value per share, buy-below price, upside/downside, valuation method, confidence, and source dates.
Use DCF/owner earnings for durable businesses, asset value for deep value, sum-of-the-parts for conglomerates, event-weighted value for special situations, and normalized cycle value for cyclicals.
If a valuation is unreliable, mark it "inconclusive" and explain what evidence is missing.

Classify the opportunity.

Compounder: high-quality business that can compound intrinsic value.
Deep value: statistically cheap with downside protection.
Special situation: catalyst-driven or event-driven asymmetry.
Macro/reflexive: driven by liquidity, policy, commodity cycle, currency, or narrative feedback loops.
Quant candidate: survives factor, base-rate, and anomaly checks.
Watchlist: good but no margin of safety.
Reject: unclear economics, weak evidence, poor risk/reward, or unacceptable governance.

Write the report.

Include executive summary, methodology, ranking table, investor-model matrix, valuation table, company sections, risks, falsification triggers, and sources.
For each serious candidate, include primary, secondary, and opposing investor lenses, the calculations used by those lenses, and the facts that support or weaken each lens.
Avoid hype. Separate facts, assumptions, and judgments.
State uncertainty clearly and include what would change the conclusion.

Verify quality before finalizing.

Check ticker, exchange, currency, fiscal period, share count, one-off items, leverage, source dates, and whether current price data is fresh.
Do not fabricate citations or fill missing data with invented numbers.

Required Output for Ranked Lists

Every Top-N or ranked candidate answer should include:

Column

Requirement

Rank

Ordered by risk-adjusted opportunity, not only quality.

Ticker/company

Include exchange and currency when useful.

Current price

Include source and access date.

Intrinsic value or fair value

Include method and confidence.

Buy-below price

Apply explicit margin of safety.

Upside/downside

Use current price versus intrinsic/fair value.

Best-fit investor lenses

Name the strongest mental models.

Composite score

Use the seven-axis framework.

Thesis

One concise, evidence-backed reason to care.

Main risk

One concise reason the thesis could fail.

Supporting Files

Investor models: references/investor-mental-models.md

Detailed investor mental models: references/great-investor-mental-models.md

Evaluation framework: references/evaluation-framework.md

Valuation methods: references/valuation-playbooks.md

Data sourcing: references/data-source-guide.md

Report writing: references/report-writing-guide.md

Red flags: references/risk-red-flags.md

Report template: templates/research_report_template.md

Company card schema: templates/company_card_schema.json

Korean user guide: docs/README_KO.md

Installation guide: docs/INSTALL.md

Korean installation guide: docs/INSTALL_KO.md

Korean mental-model guide: docs/GREAT_INVESTOR_MENTAL_MODELS_KO.md

Candidate scoring script: scripts/score_candidates.py

Report builder script: scripts/build_report_from_cards.py

Source validator: scripts/validate_sources.py