Company Teardown

⚠️ **Finance warning / Avertissement finance** : informational only, not investment advice. A teardown organizes what disclosure supports, it does not validate a business: inferred figures can be wrong, a well understood company can still be a bad investment, there is a real risk of loss, and no brief here is a promise of return. Curated by Skill Harbor: the company teardown skill at the root of andreworia/claude-finance-skills. Use it at the start of any sell-side or buy-side mandate, before drafting a CIM section, pitching coverage or building a financial model. It reverse engineers the target from disclosed and inferred data through six lenses that every teardown must cover: revenue, cost, capital, concentration, competitive position and governance. The workflow builds the revenue decomposition first (and warns against accepting management segmentation, which is designed for compensation, not analysis, so it re-segments along the dimension that predicts behavior), decomposes the cost structure into fixed, variable and stranded costs, maps the unit economics appropriate to the business model, builds the cash flow conversion view with an EBITDA to free cash flow bridge that separates maintenance capex from growth capex (headline EBITDA blindness and missing maintenance capex are named failure modes), runs concentration analysis across customer, supplier, geography and channel with magnitude and trend, tests the competitive position with a margin defense thesis backed by three to five years of margin and share evidence, and reads ownership, governance and the management incentive structure, because a teardown that ignores the shareholder math produces processes that cannot clear approval thresholds. It closes by synthesizing everything into a one page brief with the investment thesis and a list of disclosure gaps to address in management meetings or the data room, and it insists that every material claim be sourced to primary disclosure such as a filing, a transcript or a contract. From the andreworia/claude-finance-skills repository (MIT). Honest caveats: the filings, transcripts and operating data are yours to supply and to verify; inferred data is labelled as inference inside the method, and gaps stay gaps until primary sources close them. Skill Harbor never reviews the code, review it yourself before use.