What problem does it solve? Valuations often start from a driver template and grow a story afterwards, producing rationalizations no downstream check can detect. This Skill reverses that order: it writes the business story first, grades every claim by evidence, and converts each claim into exactly one DCF driver so the forecast is grounded in an auditable narrative. ## Core Features & Use Cases - Narrative-to-driver mapping: Produces narrative.md (the prose case) and drivers.json (the dcf-valuation-engine payload), with a one-claim-one-driver rule and a claim ledger graded probable, plausible, or possible. - Constraint and screen enforcement: Honors classification constraints such as no-standard-growth-model, require-failure-probability, and no-perpetual-growth-above-riskfree, and runs impossible/implausible/improbable screens plus an aggregation test. - Counter-narratives and probe diagnostics: Builds higher and lower alternative stories, then probes the payload through the DCF engine to read terminal value share, implied marginal return on capital, and terminal excess return. - Use Case: As the narrative stage of a valuation team run, it receives classification and financial inputs from an orchestrator and hands defensible growth, margin, reinvestment, and failure assumptions to the intrinsic-valuation stage. ## Quick Start Ask the valuation orchestrator to run the narrative stage for a company so it writes narrative.md and drivers.json from the classification and financial inputs.
