Install
$ agentstack add skill-deviesvi-skills-global-stock-value-analysis ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Global Stock Value Analysis
Unified workflow for value-investing oriented stock analysis across US stocks, China A-shares, HK stocks, ADRs, and mixed portfolios. The default lens is long-term business ownership: fundamentals, valuation, margin of safety, and risk first; market news is incorporated when it changes facts, probabilities, catalysts, or downside risk.
When To Use
Use this skill when the user asks to:
- analyze a listed company or ticker
- screen stocks by value, quality, dividend, growth, or balance-sheet metrics
- compare several stocks or companies across the same or different markets
- estimate intrinsic value, fair value range, or margin of safety
- review fundamentals, financial health, accounting red flags, or business quality
- review market news, earnings events, regulatory changes, macro shocks, or company-specific catalysts
- produce an investment memo, watchlist, or portfolio-level research note
This skill is research support, not personalized financial advice. Use current data and state data dates clearly. For any current-market view, verify price, filings, and news recency before analysis; stale data must be labeled and should not drive a confident conclusion.
Investment Philosophy
Prefer value-investing reasoning over short-term market commentary:
- Treat a stock as fractional ownership of a business.
- Start with durable earnings power, cash generation, balance-sheet resilience, and capital allocation.
- Require a margin of safety before calling valuation attractive.
- Use news to update the thesis, not to chase price action.
- Distinguish permanent impairment from temporary volatility.
- Admit uncertainty and show the assumptions that would change the conclusion.
- Stay inside the analyst's circle of competence. If the business, accounting, regulation, or technology is not understandable, mark it as "too hard" rather than forcing a view.
- Prefer great businesses with durable economics and fair prices over merely cheap businesses with weak economics.
- Apply Buffett-style owner earnings when reported earnings and free cash flow differ materially.
- Apply Munger-style inversion: ask what could permanently destroy the investment case, then test whether those failure modes are plausible.
- Apply Duan Yongping's lens: right business, right people, right price. Treat business model and corporate culture as primary, and price as the discipline that protects against overpaying.
Market Routing
First classify the security and choose data sources:
| Market | Examples | Preferred data sources | | --- | --- | --- | | US stocks | AAPL, MSFT, NVDA, BRK.B | SEC filings, company IR, Yahoo Finance, Nasdaq, Macrotrends, Koyfin/TradingView pages, reputable financial news | | China A-shares | 600519, 000858, 300750 | akshare, exchange filings, company announcements, financial statements, Chinese financial portals | | HK stocks | 0700.HK, 9988.HK | HKEX filings, company IR, Yahoo Finance, exchange/company announcements | | ADRs/global | BABA, TSM, ASML | Home-market filings, SEC ADR filings if applicable, company IR, exchange data |
If the market is ambiguous, infer from ticker format and company name. Ask one concise clarification only when ambiguity would materially change the analysis.
Core Workflow
- Define task type: screening, single-stock analysis, comparison, valuation, risk review, or full investment memo.
- Gather current price, market cap, filings/financials, key ratios, segment/business context, recent news, and peer data.
- Run the recency gate: timestamp market data, identify latest filing periods, check latest company announcements, and scan recent news before making valuation or risk claims.
- Verify recency. Prefer latest annual report plus latest quarterly/interim report. Mention fiscal period, report date, source, and retrieval date.
- Analyze business quality: moat, industry structure, unit economics, management/capital allocation, cyclicality, and regulation.
- Analyze financials: revenue growth, margins, ROE/ROIC, cash conversion, free cash flow, leverage, liquidity, dilution, dividends/buybacks.
- Apply the value-investor gate: circle of competence, business model quality, management/culture, moat durability, owner earnings, and margin of safety.
- Analyze valuation: historical multiples, 5/10/20-year P/E percentile when P/E is meaningful, peer multiples, DCF/DDM where appropriate, owner earnings or FCF yield, and margin of safety.
- Analyze news and events: classify each item as thesis-changing, catalyst, risk signal, sentiment/noise, or data point needing confirmation.
- Check risks and red flags: accounting quality, debt maturity, customer concentration, governance, related-party transactions, regulatory exposure, cyclicality, and recent adverse news.
- Produce a clear conclusion with thesis, key numbers, valuation range, watch items, and confidence level.
Report Type Routing
Choose one report format before drafting:
| User request | Required format | | --- | --- | | "Analyze X", "怎么看X", "分析X", or other broad single-stock asks | Quick Stock Note | | "Full report", "deep dive", "完整报告", or "深度分析" | Full Investment Memo | | "Latest news", "最近发生什么", "催化", or news-driven asks | News-Aware Value Memo | | "Compare", "对比", "X vs Y", or peer ranking asks | Comparison Output | | "估值", "值不值得买", "fair value", or entry-price asks | Valuation-Focused Note | | "风险", "暴雷", "财务问题", or downside-only asks | Risk Review |
Mandatory Output Contract
For any stock analysis, the final answer MUST follow references/report-template.md.
Before writing the final answer:
- Select exactly one report format from Report Type Routing.
- Read
references/report-template.md. - Preserve the selected template's section order. Translate/localize required headings to match the target output language while preserving their meaning.
- Do not omit required sections. If data is unavailable, write "未验证/暂无可靠数据" in that section and lower confidence.
- Put conclusion first for single-stock notes unless the selected template explicitly says otherwise.
- Separate facts, estimates, and judgment instead of blending them in the same paragraph.
- For single-stock reports, explicitly apply the Buffett/Munger/Duan gates: durable business quality, trustworthy management/culture, owner earnings or cash generation, inversion risk, and right price/margin of safety.
References
Read only what is needed:
references/framework.mdfor the unified value-investing checklist and scoring rubric.references/investment-philosophy.mdfor Buffett, Munger, and Duan Yongping principles translated into actionable analysis gates.references/market-data.mdfor market-specific data gathering guidance.references/news-events.mdfor market news, catalysts, and event triage.references/report-template.mdfor mandatory report formats. Read it before every final stock-analysis answer.
Output Rules
- Always separate facts, estimates, and judgment.
- Always include source dates for current market data and financial periods.
- For price-sensitive conclusions, include retrieval date/time and market status when available: open, closed, pre-market, after-hours, suspended, or holiday.
- If current price, filings, or material news cannot be verified, downgrade confidence and label the analysis as stale-data or limited-data.
- Always separate durable thesis items from short-term news flow.
- Avoid overprecision in valuation. Use ranges and key assumptions.
- For DCF/DDM, state discount rate, terminal growth, normalized FCF/dividend, and sensitivity.
- For P/E percentile, choose the most reasonable 5/10/20-year history for the specific company and explain why that horizon fits its listing age, business maturity, cycle length, accounting comparability, and data reliability.
- If P/E is distorted by one-off or non-operating items such as tax provisions, impairment, litigation, restructuring charges/gains, subsidies, or asset disposals, normalize earnings first, use adjusted P/E percentile when defensible, and explicitly warn the user about the adjustment and residual uncertainty.
- For cross-market comparisons, normalize currency, accounting standards, fiscal periods, and one-off items.
- If data is unavailable or stale, say so and continue with a clearly marked limited analysis.
Final Self-Check
Before final response, verify:
- Current price/date is stated, or marked unavailable.
- Latest reporting period is stated, or marked unavailable.
- The selected template's required sections are present, in order, and their headings match the target output language.
- Facts, estimates, and judgment are separated.
- Sources are listed with dates where possible.
- Any unavailable data is clearly labeled and reflected in confidence.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: DeviesVi
- Source: DeviesVi/skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.