ux-research

Router skill. Resolve topic, then Read EXACTLY ONE playbook from routes:. Never load the whole cluster — one route per invocation keeps context lean. [INFERENCIA]

When to use

A UX/discovery task that maps to one route. NOT for build/UI work (frontend skills), product strategy, or analytics instrumentation. [INFERENCIA]

Inputs → Outputs

• In: topic (required), depth (quick|deep), the request + any artifacts. [DOC]
• Out: the routed playbook's deliverable (interview guide, survey, test plan, prototype spec) with evidence tags and an explicit next step. [DOC]

Route selection

• user-research — generative: interviews, contextual inquiry, "why/what do users need". [DOC]
• survey-design — quantitative attitudes/segmentation at scale. [DOC]
• user-testing — evaluate an existing flow/prototype: usability, task success. [DOC]
• prototyping — produce a testable artifact before build. [DOC] Two topics plausible → ask one disambiguating question; never run two routes. [INFERENCIA]

Depth & spine

deep → apply playbook exhaustively, verify each step; quick → essentials, single pass. Spine: Discover → Analyze → Execute → Validate. [DOC]

Validation gate (before "done")

• Exactly one playbook was read and followed. [DOC]
• Output respects constitution v6.0.0 + script-first rule. [CONFIG]
• Every non-obvious claim carries one Alfa-core tag (references/verification-tags.md). [DOC]
• Each [SUPUESTO] is paired with a verification step. [DOC]
• Score against assets/quality-rubric.json (universal + route criteria). [CONFIG]

Anti-patterns

Loading >1 route; inventing a topic outside the enum; skipping tags; emitting a deliverable with no next step; doing the build instead of the research. [INFERENCIA]