What problem does it solve? Finished valuation models often contain double-counted risks, route violations, and unsupported terminal assumptions that pass internal consistency checks. This Skill provides the adversarial review stage that systematically attacks a completed DCF, relative, or deal analysis and converts every defect into an actionable finding. ## Core Features & Use Cases - Eleven-step attack sequence: Runs route-conformance, bias, single-charge-per-risk, growth-cost, terminal-value, equity-bridge, implied-expectations, comparable-set, seven-sins, traceability, and uncertainty attacks in a fixed order. - Mechanical validation gate: Executes the valuation-consistency-checks validator first and converts ERROR/WARN output into severity-ranked findings. - Structured findings output: Writes challenge.json with six-key findings (id, severity, target_stage, claim, evidence, suggested_fix) plus a five-section challenge.md prose companion. - Use Case: After an orchestrated DCF run values a company, this Skill recomputes what growth costs, reverse-solves the market price against the analyst's drivers, and reports that country risk was charged twice — once as a premium and once as a haircut. ## Quick Start Ask the agent to run the valuation critic stage against the mandate, classification, forecast, and DCF result artifacts and produce challenge.json and challenge.md.
