What problem does it solve? Analysts working on investment research often re-derive the same metrics from raw evidence for every report. This Skill transforms a curated evidence ledger into standardized signal cards before analyst fan-out, so downstream experts consume consistent, traceable signals instead of re-interpreting raw sources. ## Core Features & Use Cases - Signal Card Generation: Reads evidence-plan.md and evidence-ledger.md from the active workspace and writes signal-cards.md with subject, geography, period, input evidence IDs, calculation, confidence, and caveats. - Claim Boundary Enforcement: Keeps claim boundaries visible next to each signal and writes low-confidence or blocked signals instead of inventing metrics when evidence is insufficient. - Use Case: Before running multiple analyst roles on a company like CSTM, run this Skill to convert the collected evidence ledger into reusable signal primitives (e.g., revenue trend, margin direction) that each analyst cites rather than recomputing. ## Quick Start Generate signal cards from the evidence ledger in the active workspace before starting the analyst fan-out.
