본문으로 건너뛰기
cccskills
무료GitHub에 공개

nw-discovery-methodology

Question-first approach to understanding user journeys. Load when starting a new journey design or when the discovery phase needs deepening.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(1)

  • SKILL.md4.7 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Discovery Methodology

Discover journeys through deep questioning before any sketching. The sketch is proof of understanding, not the starting point.

Session Flow

Phase 1: Goal Discovery (First 5-10 minutes)

Focus: What is the user trying to accomplish?

Questions:

  • "What's the ultimate goal you're trying to achieve?"
  • "What triggers this journey? When does a user start this?"
  • "How will you know when you've succeeded?"
  • "What's the happy path in your mind?"

Draft Sketch (after first 3 answers)

After collecting goal, trigger, and success criteria, output a rough draft sketch showing what the journey might look like. Mark unknowns with ???. Purpose: make value visible immediately so the user sees progress, not just questions.

[Trigger: ???] → [Step 1: ???] → [Step 2: ???] → [Goal: {stated goal}]
  Feels: ???       Sees: ???       Sees: ???       Feels: {success criteria}
  Artifacts: ???   Artifacts: ???  Artifacts: ???

This is a working hypothesis, not a commitment. Update it after each phase as understanding deepens. The sketch gives the user something concrete to react to — "no, step 2 happens before step 1" is more productive than abstract discussion.

Phase 2: Mental Model (10-20 minutes)

Focus: What does the user EXPECT to see?

Questions:

  • "Walk me through step by step -- what do you type, what do you see?"
  • "At this step, what information appears on screen?"
  • "What would you need to see to feel confident?"
  • "Where do you think this data comes from?"
  • "What's your mental model of how this works?"

Phase 3: Emotional Journey (5-10 minutes)

Focus: How should the user FEEL?

Questions:

  • "How should you feel at the start? Anxious? Curious? Confident?"
  • "What's the emotional arc -- where's the peak tension?"
  • "How should you feel when it's done?"
  • "Where might you feel frustrated or lost?"
  • "What would make this feel satisfying vs frustrating?"

Phase 4: Shared Artifacts (5-10 minutes)

Focus: What data is shared across steps?

Questions:

  • "What information appears in multiple places?"
  • "Where does the version number come from? Where is it shown?"
  • "If this path changes, what else breaks?"
  • "Who owns this piece of data?"
  • "What paths or URLs are reused across steps?"

Phase 5: Error Paths (5 minutes)

Focus: What could go wrong?

Questions:

  • "What's the most likely failure?"
  • "What should the user see when it fails?"
  • "How does the user recover?"
  • "What would a helpful error message look like?"

Phase 6: Integration Points

Focus: How do steps connect?

Questions:

  • "What did the previous step produce that this step needs?"
  • "What does this step produce that the next step needs?"
  • "Are there any hidden dependencies between steps?"
  • "What external systems or files does this touch?"

Phase 7: CLI UX Specifics

Focus: What commands and output does the user expect?

Questions:

  • "What command would you naturally type for this?"
  • "What flags or options do you expect?"
  • "How verbose should the output be by default?"
  • "Should there be a --dry-run option?"

Question Format

Use AskUserQuestion with structured options:

  • 2-4 concrete options based on design methodology
  • Options represent real design alternatives
  • Include an "Other" option for open-ended input
  • Each option includes its design implication

Sketch Readiness Criteria

Ready to sketch ONLY when all can be answered:

  • Complete happy path described (no "and then something happens")
  • Each step has expected output defined
  • Emotional arc is explicit and coherent
  • Shared artifacts identified with sources
  • At least major error paths acknowledged

If ANY criterion is unclear -- ask more questions.

When to Question

Always ask first when:

  • User requests a sketch -- ask before sketching
  • User describes a feature -- ask to understand journey context
  • User mentions a command -- ask what they expect to see

Continue asking until:

  • User can describe complete happy path without gaps
  • All shared artifacts identified with sources
  • Emotional arc is explicit and coherent
  • Error paths are at least acknowledged

Anti-Patterns

  • Jumping to sketching before understanding mental model
  • Assuming you know what the user expects
  • Filling gaps with your own assumptions
  • Skipping emotional journey questions
  • Ignoring shared artifact tracking

Instead:

  • Ask one more question when uncertain
  • Reflect back understanding for user validation
  • Make the user articulate their mental model explicitly
  • Map emotional states at every step
  • Document every ${variable} and its source

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

nWave-ai/nWave6212026년 9월 16일 업데이트

Review dimensions for validating agent quality - template compliance, safety, testing, and priority validation

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

nWave-ai/nWave6212026년 9월 16일 업데이트

Review dimensions for acceptance test quality - happy path bias, GWT compliance, business language purity, coverage completeness, walking skeleton user-centricity, priority validation, observable behavior assertions, traceability coverage, and walking skeleton boundary proof

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

nWave-ai/nWave6212026년 9월 16일 업데이트

Detailed 5-phase workflow for creating agents - from requirements analysis through validation and iterative refinement

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

nWave-ai/nWave6212026년 9월 16일 업데이트

5-layer testing approach for agent validation including adversarial testing, security validation, and prompt injection resistance

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

nWave-ai/nWave6212026년 9월 16일 업데이트

Architectural style selection decision matrices, trade-off analysis, structural enforcement rules, and combination patterns. Load when choosing or evaluating architecture styles.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

nWave-ai/nWave6212026년 9월 16일 업데이트

nWave-ai의 스킬 모두 보기

이 스킬의 문제 신고