dsh-plugin-user-research

The hardest part of user research is rarely a lack of data. It is the gap between a pile of transcripts, open-ended survey responses and support tickets, and a conclusion you can actually defend in a review. Let an LLM summarize and it will happily invent pain points the user never said. dsh-plugin-user-research closes that gap with a fixed research framework that synthesizes raw, verbatim user input into structured personas, evidence-backed pain points, and ranked opportunities — without ever fabricating a sentence the user did not write.

Three principles drive the output. A six-dimension theme taxonomy (efficiency, usability, cost, missing features, stability, support) automatically classifies scattered quotes. An anti-hallucination constraint means pain points are always verbatim quotes with a severity tag, never paraphrased into "the user believes …", and low-sample results carry an explicit low-confidence flag. Opportunity scoring ranks candidates by mention frequency and intensity of expressed pain, so you can triage directions without guessing. Four reusable research skills and two markdown templates (interview guide, synthesis canvas) cover the full workflow from raw notes to a decision-ready brief.

If you are a product manager synthesizing ten interview transcripts, a UX researcher pulling high-frequency issues from usability logs, an entrepreneur mining open-ended survey responses and competitor reviews, or a customer-success team clustering complaints to set improvement priority — any one of these roles that holds a stack of user voice and wants a clear, evidence-based answer to "who hurts, and is it worth doing" is exactly who this plugin is built for.