--- name: company-valuation description: "Performs multi-model company valuation. Selects the best 3-4 models from a library of 22, sources data from PitchBook, FactSet, SEC EDGAR, and the web, runs the calculations, and delivers an institutional-quality report with executive summary, results table, and a full section per model. Use whenever a user asks to value, price, or analyse a company." metadata: display_name: "Company Valuation" short_description: "Select the best 3-4 valuation models, fetch live data, run the analysis, and deliver an institutional-quality valuation report." default_prompt: "Value Apple (AAPL) as a long-term buy-and-hold investment using the most appropriate models." icon_path: "assets/icon.png" --- # Company Valuation Analyst ## Overview This skill performs end-to-end company valuation analysis. It reads `references/company_valuation_models.md` to understand the library of 22 models, then follows a strict five-phase workflow: 1. **Parse** — understand the company and the investment scenario 2. **Select** — choose the 3–4 models best suited to the company and scenario 3. **Source** — retrieve all required financial data from live connectors and filings 4. **Calculate** — run every selected model using the formulas in the reference 5. **Report** — produce a structured markdown report --- ## Invocation ``` /company-valuation-analyst — ``` **Examples:** - `/company-valuation-analyst AAPL — long-term buy-and-hold equity investment` - `/company-valuation-analyst Microsoft — strategic acquisition analysis` - `/company-valuation-analyst a mid-market SaaS company with $50M ARR — Series C funding round` - `/company-valuation-analyst a US utility company — DCF for regulated asset base` - `/company-valuation-analyst Tesla — PE-style LBO feasibility check` --- ## STRICT WORKFLOW (follow every phase in order) --- ### PHASE 1 — PARSE & CLARIFY 1. Extract the **company identifier** (ticker, name, or description) and the **scenario** from the user's input. 2. Identify the following **scenario dimensions** — these drive model selection: - **Purpose**: Buy-side equity / Sell-side M&A / PE/LBO / VC / Credit / Internal planning - **Stage**: Early-stage / High-growth / Mature / Distressed / Conglomerate - **Sector**: Technology / Financials / Industrials / Energy / Healthcare / Consumer / Real estate / Other - **Cash flow status**: Pre-revenue / Pre-profit / Profitable and stable / Highly levered - **Dividend status**: Pays dividends / Does not pay dividends 3. If the company or ticker is ambiguous, use `web_search` to confirm the correct entity before proceeding. --- ### PHASE 2 — MODEL SELECTION Read `references/company_valuation_models.md` to review all 22 models. Select **3 to 4 models** using the decision matrix below. #### Model Selection Decision Matrix | Scenario Dimension | Strongly Suggests | |---|---| | Dividend-paying, mature, stable growth | Gordon Growth Model (GGM) or Two-Stage DDM | | Profitable, predictable cash flows, any stage | DCF (FCFF) — almost always include this | | M&A / acquisition analysis | Comparable Company Analysis + Precedent Transactions | | PE or LBO context | LBO Model + EV/EBITDA | | Pre-profit / pre-revenue tech or SaaS | EV/Revenue + VC Method | | Financial institution (bank, insurer) | Residual Income Model + P/B | | REIT / investment fund / holding company | NAV + DDM | | Conglomerate / multi-division company | Sum-of-the-Parts (SOTP) | | Distressed / credit analysis | Liquidation Value + EV/EBITDA | | Biotech / pharma with pipeline optionality | DCF + Real Options + EV/Revenue | | Resource / mining / oil & gas | DCF + NAV + Real Options | #### Mandatory defaults: - **Always include DCF (FCFF)** unless the company is pre-revenue/pre-profit OR a financial institution. - **Always include at least one market-based (relative) model** (Trading Comps or a multiple) as a cross-check. - **Never select more than 4 models** — prioritise depth over breadth. #### For each selected model, document upfront: - Why this model was chosen for this specific company and scenario - What data inputs are required - Which data sources will be used to obtain them --- ### PHASE 3 — DATA SOURCING Source all required inputs for every selected model. Use the priority order below. **Document every data point with its source** — this populates the References section of the report. #### Source Priority Order **1. PitchBook** (only if the user has a PitchBook connector or licence available; otherwise skip to the next source) - Use for: Company financials, deal comps, transaction multiples, cap table, funding rounds, revenue estimates for private companies - Use whichever PitchBook connector or tool is available in the session; do not assume one exists **2. FactSet** (only if the user has a FactSet connector or licence available; otherwise skip to the next source) - Use for: Public company financials, consensus estimates, trading multiples, peer group data, EV/EBITDA comps, cost of capital (WACC components) - Use whichever FactSet connector or tool is available in the session; do not assume one exists **3. SEC EDGAR** (via `/sec-edgar` skill) - Use for: Public company 10-K/10-Q filings, XBRL financial statements, revenue, EBITDA, net income, capex, D&A, balance sheet items - Invoke as: use the `sec-edgar` skill and follow its workflow to extract the latest 10-K or 10-Q for the target ticker **4. Web Search** - Use for: Beta estimates, risk-free rate (current 10-year Treasury yield), equity risk premium, analyst consensus estimates, sector EV/EBITDA multiples, dividend history, recent news - Always check publication dates; use 2026 figures where available **5. Web fetch / Manual** - Use for: Specific data points from investor relations pages, proxy statements, earnings call transcripts #### Data Checklist by Model **DCF (FCFF):** - [ ] Revenue (last 3 years actuals + 3-5 year forecasts) - [ ] EBITDA margin (historical + projected) - [ ] D&A, Capex, change in working capital - [ ] Tax rate (effective) - [ ] WACC components: Risk-free rate, equity risk premium, beta, cost of debt, capital structure - [ ] Terminal growth rate (long-run GDP or sector growth) **Gordon Growth Model / DDM:** - [ ] Annual dividend per share (most recent) - [ ] Dividend growth rate (5-year CAGR or analyst forecast) - [ ] Cost of equity (from CAPM: Rf + beta × ERP) **EV/EBITDA Multiple:** - [ ] Target company LTM EBITDA - [ ] Peer group EV/EBITDA median (5–8 comparable companies) - [ ] Net debt (total debt minus cash) **Comparable Company Analysis:** - [ ] Target: Revenue, EBITDA, EBIT, Net Income, EPS - [ ] Peer group: Same metrics + current market cap + net debt for each peer - [ ] Calculate: EV/Revenue, EV/EBITDA, P/E for each peer; take median **LBO Model:** - [ ] Entry EV/EBITDA multiple (from precedent transactions) - [ ] LTM EBITDA - [ ] Revenue growth and EBITDA margin projections (5 years) - [ ] Leverage assumption (Debt/EBITDA at entry, typically 4–6x) - [ ] Interest rate on debt - [ ] Exit EV/EBITDA multiple assumption - [ ] Target IRR (typically 20%+) **Residual Income / P/B:** - [ ] Book value of equity (most recent balance sheet) - [ ] ROE (net income / avg. book equity) - [ ] Cost of equity - [ ] Long-run growth rate **NAV:** - [ ] Fair market value of all assets (use cap rates for real estate, DCF for operating assets) - [ ] Total liabilities **Sum-of-the-Parts:** - [ ] Segment revenue and EBITDA - [ ] Appropriate peer multiple per segment - [ ] Consolidated net debt **VC Method:** - [ ] Projected revenue or EBITDA at exit (Year 5–7) - [ ] Exit multiple (sector EV/Revenue or EV/EBITDA) - [ ] Target IRR (typically 30–50%) - [ ] Investment amount **Real Options:** - [ ] PV of project/asset cash flows (S) - [ ] Investment required (K) - [ ] Time to decision (T) - [ ] Risk-free rate (r) - [ ] Asset volatility (sigma — use comparable company stock volatility or project cost volatility) **Liquidation Value:** - [ ] Balance sheet: all asset categories with book values - [ ] Apply recovery rate assumptions by asset class - [ ] Total liabilities --- ### PHASE 4 — CALCULATION For each selected model: 1. Read the exact formula from `references/company_valuation_models.md` 2. Populate all inputs with sourced data 3. Execute the calculation step by step (show arithmetic) 4. Derive the **implied enterprise value** and **implied equity value per share** (or per unit as appropriate) 5. Perform **sensitivity analysis** for 1–2 key inputs (e.g., for DCF: vary WACC ±1% and terminal growth ±0.5%; for multiples: use 25th/median/75th percentile of peer group) 6. State the **valuation range** (low / base / high) for each model Use Python (`execute_command`) when calculations are complex (e.g., DCF with multi-year projections, LBO IRR, Black-Scholes for real options). Save calculation scripts to `scripts/` directory during runs for transparency. --- ### PHASE 5 — REPORT GENERATION Produce a single structured markdown report using the exact format below. Save it as `valuation_report__.md` in the workspace. --- ## OUTPUT REPORT FORMAT (MANDATORY) ```markdown # Valuation Analysis: [Company Name] ([Ticker]) **Date:** [Date] **Scenario:** [Investment scenario description] **Analyst:** Company Valuation Analyst Skill --- ## Executive Summary [3–5 paragraph narrative covering: - Company overview and why this scenario was chosen - Which models were selected and why they are the most appropriate for this company/scenario - Key findings: implied valuation range across all models - Investment conclusion / key risks] --- ## Valuation Summary Table | Model | Category | Implied EV ($M) | Implied Equity Value ($M) | Implied Price / Share | Weight | Notes | |-------|----------|:-:|:-:|:-:|:-:|-------| | [Model 1] | [Income/Market/Asset] | $X – $Y | $A – $B | $P – $Q | [%] | [1-line note] | | [Model 2] | ... | ... | ... | ... | ... | ... | | [Model 3] | ... | ... | ... | ... | ... | ... | | [Model 4 if applicable] | ... | ... | ... | ... | ... | ... | | **Blended / Triangulated Range** | | **$X – $Y** | **$A – $B** | **$P – $Q** | 100% | | --- ## Model 1: [Model Name] ### Why This Model Was Selected [1–2 paragraphs explaining why this model is appropriate for this company and scenario, referencing the company's stage, sector, cash flow profile, and the investor's purpose.] ### Key Inputs | Input | Value | Source | Rationale | |-------|-------|--------|-----------| | [Input name] | [Value with units] | [Source: PitchBook / FactSet / SEC 10-K / Web] | [Why this value was used] | | ... | ... | ... | ... | ### Calculation [Show the formula, then substitute values step by step] **Formula:** [formula from reference] **Step-by-step:** ``` [arithmetic, shown clearly] ``` **Sensitivity Analysis:** | [Key Input] ↓ / [Other Input] → | [Low] | [Base] | [High] | |---|---|---|---| | [Low] | $X | $Y | $Z | | [Base] | ... | ... | ... | | [High] | ... | ... | ... | **Result:** Implied EV = $X – $Y | Implied Equity Value = $A – $B | Implied Price/Share = $P – $Q ### References - [Source 1: full citation with URL or filing date] - [Source 2: ...] --- ## Model 2: [Model Name] [Repeat same structure] --- ## Model 3: [Model Name] [Repeat same structure] --- ## Model 4: [Model Name, if applicable] [Repeat same structure] --- ## Methodology Notes & Limitations [Paragraph covering: - Key assumptions that most influence results - Data gaps or limitations encountered - Models that were considered but excluded, and why - Caveat about forward-looking assumptions] --- ## Data Sources | # | Source | Type | Data Retrieved | Date Accessed | |---|--------|------|----------------|---------------| | 1 | PitchBook | Financial data platform | [list data points] | [date] | | 2 | FactSet | Financial data platform | [list data points] | [date] | | 3 | SEC EDGAR (10-K/10-Q) | Regulatory filing | [list data points] | [date] | | 4 | Web Search | Market data | [list data points] | [date] | ``` --- ## Constraints & Failure Handling - **Private company — no public filings**: Use PitchBook for financials; use VC Method or EV/Revenue multiples; skip DCF if no historical cash flow data available. - **No dividend history**: Do not use DDM variants; substitute DCF or residual income. - **PitchBook/FactSet not available or returns no data**: Fall back to SEC EDGAR + web search; note the gap in the report. - **Insufficient peer group data**: Use published sector reports from Damodaran (NYU Stern) or industry research; cite clearly. - **Calculation fails or data missing**: Show partial calculation, clearly flag the missing input, and state the range is indicative only. - **Maximum 4 models**: If a 5th model would add meaningful perspective, mention it briefly in Methodology Notes instead. --- ## Reference Files - `references/company_valuation_models.md` — Full library of 22 models with formulas, complexity ratings, and usage guidance - `scripts/model_selector.py` — Model selection logic helper - `scripts/valuation_engine.py` — Calculation functions for all models - `scripts/report_generator.py` — Report assembly and formatting helper --- ## Example Run **Input:** `/company-valuation-analyst MSFT — strategic acquisition analysis by a technology acquirer` **Phase 1 parse:** - Company: Microsoft Corporation (MSFT), large-cap technology - Scenario: M&A acquisition → purpose is to determine fair acquisition price - Stage: Mature, highly profitable - Sector: Technology (enterprise software) - Cash flow: Strongly FCF-positive - Dividends: Yes (growing) **Phase 2 model selection:** 1. DCF (FCFF) — intrinsic value anchor; MSFT has highly predictable FCF 2. Comparable Company Analysis (EV/EBITDA, P/E) — market reference; large peer group exists 3. Precedent Transaction Analysis — what large-cap software acquisitions have traded at 4. Two-Stage DDM — MSFT pays dividends with consistent growth; cross-check on equity value **Phase 3 data sourcing:** FactSet for peer multiples and financials → SEC EDGAR 10-K for FCFF inputs → web search for current WACC components → PitchBook for any M&A transaction comps **Phase 4 calculation:** Run each model, produce EV range **Phase 5 report:** Deliver full markdown report with all sections