AgentSkillsCN

ux-research-synthesis

针对将定性与定量研究结果转化为优先级设计行动的UX研究综合工作流程。当UX、交互、视觉或设计治理相关产物为主要交付成果时使用;切勿用于后端数据模型或部署流水线的决策。

SKILL.md
--- frontmatter
name: ux-research-synthesis
description: "UX research synthesis workflow for translating qualitative and quantitative findings into prioritized design actions. Use when UX, interaction, visual, or design-governance artifacts are the primary deliverable; do not use for backend data-model or deployment pipeline decisions."

UX Research Synthesis

Trigger Boundary

  • Use when research data exists but design implications are unclear.
  • Do not use for running stakeholder interviews; use stakeholder-interview.
  • Do not use for direct interface handoff packaging; use figma-handoff.

Goal

Turn research evidence into high-confidence, actionable design direction.

Shared Design Contract (Canonical)

  • Use ../design-principles/references/design-governance-contract.md as the single schema and gate source.
  • Track synthesis packages with UX-RSR-* IDs.
  • Run machine validation: python3 ../design-principles/scripts/validate_design_contract.py --manifest <path/to/manifest.json>.

Inputs

  • Research transcripts, notes, analytics, and observations
  • Product goals and known design hypotheses
  • Segment definitions and sampling limitations

Outputs

  • Synthesized findings package with UX-RSR-* ID
  • Prioritized design implications and risk notes
  • Confidence level and evidence traceability map
  • Privacy evidence package for research handling decisions

Workflow

  1. Normalize research inputs and remove duplicate signals.
  2. Cluster findings by user goal and failure pattern.
  3. Distinguish observed facts from hypotheses.
  4. Map findings to design implications and priority.
  5. Publish synthesis with confidence, limitations, and privacy evidence.

Quality Gates

  • Every implication links back to evidence.
  • Confidence level and sampling limits are explicit.
  • Contradictory signals are surfaced, not hidden.
  • Privacy Reviewer approval is always present.
  • Contract validation passes including privacy evidence requirements.

Failure Handling

  • Stop when evidence traceability is missing.
  • Stop when Privacy Reviewer approval is missing.
  • Escalate when synthesis relies on non-representative samples.