company-valuation

This skill estimates the intrinsic value of a public company using three complementary methods and triangulates them into an implied share price with upside or downside versus the current market price. It solves the problem of answering valuation questions rigorously rather than from memory: it runs a discounted cash flow (5-year FCFF projection discounted at WACC with a terminal value), a relative valuation using peer median multiples (P/E, EV/Revenue, EV/EBITDA), and a sum-of-the-parts analysis when a company has two or more distinct reporting segments, then blends them and presents a sensitivity table and Bull/Base/Bear scenarios. Output is explicitly framed as research and educational, not financial advice.

The workflow begins with a detection step that checks runtime dependencies and data sources: it probes for yfinance (with numpy and pandas) as the primary path, falls back to a funda-data provider skill, and can pip-install yfinance when nothing is available. It fetches the 10-year Treasury yield as a live risk-free rate, defaulting to 4.5% if that fetch fails. Reference files provide detailed methodology for DCF, relative valuation, SOTP, and WACC/equity-risk-premium rates, along with sector-specific guidance (banks use DDM or excess-return models, REITs use NAV and P/FFO, cyclicals use mid-cycle normalization, and so on) and common pitfalls such as terminal-value dominance and double-counting stock-based compensation.

It targets investors, analysts, and finance-literate users who want a fair-value estimate, a DCF build, or a peer comparison for a given ticker. It runs Python for financial computation and reads market data through yfinance; it performs read-only data retrieval and local modeling, and may pip-install its dependencies.