What problem does it solve? Large sets of reviews, comments, interviews, and support conversations are hard to act on: duplicates inflate apparent consensus, minority concerns get averaged away, and requested solutions get mistaken for validated diagnoses. This Skill converts a bounded set of supplied feedback into a traceable atomic ledger, evidence-backed themes, and one provisional disposition per issue. ## Core Features & Use Cases - Atomic feedback ledger: Normalizes every source into atomic entries with stable IDs, exact wording, locators, severity, and source limitations, then deduplicates while preserving raw mention counts. - Theme and tension mapping: Clusters by root concern and affected outcome, preserving contradictions, minority audiences, and isolated high-severity findings instead of forcing consensus. - Disposition recommendations: Assigns exactly one of accept, reject, clarify, test, defer, or already-addressed per issue, with rationale, confidence, and validation actions. - Use Case: After a product launch, feed 40 app-store reviews, 12 support transcripts, and 5 user interviews into the Skill to get a deduplicated theme map showing which concerns affect which audiences and what to validate first. ## Quick Start Use the sd-feedback skill to synthesize the attached customer reviews and interview notes into themes and response dispositions for our onboarding flow.
