user-research-prior-art-survey

Surveys published user-research literature into schema-validated evidence registers and reports.

1|1|Updated May 24, 2026
One-click install
npx skills add https://github.com/bm629/agent-skills --skill user-research-prior-art-survey-bm629
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: user-research-prior-art-survey
Source: https://github.com/bm629/agent-skills/tree/main/skills/user-research-prior-art-survey
Command: npx skills add https://github.com/bm629/agent-skills --skill user-research-prior-art-survey-bm629

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Before designing an interface, teams rarely know what published research already says about their users and tasks, and ad-hoc literature searches produce unverifiable, unreproducible summaries. This Skill turns a product scope into a rigorous, schema-governed survey of published user-research evidence with validated artifacts at every stage. ## Core Features & Use Cases - Research vocabulary mapping: Derives a four-axis search protocol (user-population, task, method, component) with typed term expansions, honest provenance, and per-angle applicability verdicts. - Angle-based retrieval: Executes one named search angle across scholarly indexes (Crossref, Semantic Scholar, arXiv), practitioner corpora (NN/g, Baymard, GOV.UK), and standards-body findings, recording every query verbatim and honoring declared crawl delays. - Finding extraction and synthesis: Deep-reads one source into N finding records with rule-assigned GRADE certainty and separate transferability judgments, then synthesizes an evidence register and seven-section report. - Use Case: Before designing a mobile checkout flow, run the survey to build the vocabulary map, execute the practitioner-research angle, extract findings from admitted full-text sources, and produce an evidence register whose every claim carries a finding id, certainty level, and transferability rationale. ## Quick Start Use user-research-prior-art-survey to build the research vocabulary map for this product scope before we design the interface.

Frequently Asked Questions about user-research-prior-art-survey

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I survey published user research before designing a product interface?

Build a research vocabulary map from the product scope across four axes (population, task, method, component), then execute one search angle at a time against scholarly indexes and practitioner corpora. Each artifact is validated against a JSON schema before the next wave proceeds.

What sources does a user-research prior-art survey cover?

The registry covers Crossref, Semantic Scholar, arXiv listings, Nielsen Norman Group, Baymard's free articles, GOV.UK service manual, Europe PMC's REST API, W3C WAI, Android and Apple developer guidance, and the Stack Overflow developer survey, each with verified access terms.

Does this skill conduct user interviews or usability tests?

No. It only reads published findings from peer-reviewed work, practitioner research, and standards bodies. It runs no study of its own; conducting interviews, usability tests, or original surveys is explicitly out of scope.

How is evidence certainty graded in the survey output?

Certainty uses GRADE's four-level vocabulary (high, moderate, low, very-low) assigned by a deterministic rule from study design, sample size, and effect size. The validator re-derives the level and rejects mismatches, so it is arithmetic rather than appraisal.

Why must sources be admitted only with full text and a stated method?

A record built from an abstract is indistinguishable from one grounded in the method section, and full text without a stated method yields nothing weighable. Both conjuncts are required; everything else is recorded as unadmitted with its reason.

What happens when a source is rate-limited or unreachable during the search?

Each coverage cell records a typed status such as rate-limited, unreachable, or forbidden-by-terms with a cause, never a silent zero. A throttled shared pool like Semantic Scholar's is a normal operating condition, not an empty result set.