について
このスキルは、フックモデル(トリガー、アクション、可変報酬、投資)を用いて、習慣化する製品ループをデザインする開発者を支援します。ユーザー定着率の問題に対処する際、通知や連続記録などのエンゲージメントシステムを構築する際、あるいはサインアップ後の離脱を分析する際にご活用ください。また、倫理的評価と習慣形成のオンボーディングに関するガイダンスも含まれています。
クイックインストール
Claude Code
推奨npx skills add wondelai/skills -a claude-code/plugin add https://github.com/wondelai/skillsgit clone https://github.com/wondelai/skills.git ~/.claude/skills/hooked-uxこのコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします
ドキュメント
Hook Model Framework
Framework for building habit-forming products. Habits are not created — they are built through successive cycles through the Hook.
Core Principle
The Hook Model = a four-phase loop that connects the user's problem to your solution frequently enough to form a habit, moving usage from deliberate to automatic.
Trigger → Action → Variable Reward → Investment
↑ │
└──────────────────────────────────────┘
Scoring
Goal: 10/10. When reviewing or creating product engagement mechanics, score the loop by the four Quick Diagnostic rows (internal trigger, dead-simple action, variable reward, investment loads next trigger): each row earns 2 (fully satisfied), 1 (partial), or 0 (absent), then score = round(total / 8 × 10). Then apply the ethics gate: if the Manipulation Matrix places the product as Dealer (or it hits any "When NOT to Use" condition), cap the score at 3 regardless of mechanics. Bands: 9-10 = complete loop, internal trigger identified, ethics clear; 5-6 = loop runs but leans on external triggers or predictable rewards; <=3 = broken loop or extractive design. Always state the current score and the specific diagnostic rows blocking 10/10.
The Four Phases
1. Trigger
Core concept: The actuator of behavior. Triggers are external (environment-driven: notifications, emails, ads) or internal (emotion-driven) — and the goal is to migrate users from external to internal triggers.
Why it works: Every habit starts with a cue. External triggers get users started, but internal triggers — boredom, loneliness, uncertainty, FOMO — drive unprompted usage because the emotion itself fires before any reminder can.
Key insights:
- Map your product to the specific negative emotion it resolves (boredom, loneliness, confusion, FOMO)
- Effective external triggers are well-timed, actionable, and lead to the simplest possible next action
- If users still need external prompts after ~30 days, no internal trigger has formed
Product applications:
| Context | Application | Example |
|---|---|---|
| Onboarding | External triggers establish the first loop | Welcome email with one clear action |
| Retention | Map product to internal emotional trigger | Instagram resolves boredom; Google resolves confusion |
| Re-engagement | External triggers bridge gaps until habit forms | Push: "Your friend just posted a photo" |
Copy patterns:
- "Don't miss what happened while you were away" (FOMO trigger)
- "Your friend just..." (social trigger bridging to internal)
- "Pick up where you left off" (routine trigger)
Ethical boundary: Don't build triggers that fire on vulnerable emotional states (depression, addiction, grief) — those users can't exercise the autonomy the loop assumes.
See references/triggers.md when mapping triggers — the emotion-to-product mapping exercise and the external-to-internal transition plan.
2. Action
Core concept: The simplest behavior done in anticipation of a reward, guided by the Fogg Behavior Model: Behavior = Motivation + Ability + Trigger, all converging at the same moment.
Why it works: Increasing motivation is hard and unreliable; reducing friction (increasing ability) is easier and more effective. Every extra step, field, or decision is a drop-off point.
Key insights:
- Six elements of simplicity: time, money, physical effort, brain cycles, social deviance, non-routine
- The action is the simplest behavior in anticipation of reward — not the full task
- Hick's Law: more choices = slower decisions; reduce options to increase action rate
Product applications:
| Context | Application | Example |
|---|---|---|
| Signup flow | Minimize fields and steps | One-click Google/Apple sign-in |
| Core action | Completable in seconds | Twitter: type 280 characters and post |
| Progressive disclosure | Ask for more only after initial reward | Duolingo: play first, create account later |
Copy patterns:
- "Just one tap to..." (emphasizes simplicity)
- "No credit card required" (money/risk simplicity)
- Buttons should be verbs: "Post", "Save", "Share" — not "Submit" or "Continue"
Ethical boundary: When you simplify an action, don't strip the cost or consequence with it — a one-tap purchase must still surface the price and the commitment.
See references/product-applications.md when adapting the loop to your context — action and investment patterns for B2B SaaS, e-commerce, health, and productivity tools.
3. Variable Reward
Core concept: The phase that keeps users coming back. Anticipation of reward — not the reward itself — creates dopamine, and rewards must be variable (unpredictable) to sustain engagement.
Why it works: The brain's dopamine system responds most strongly to anticipation of uncertain rewards — the slot machine effect. Three reward types — tribe (social), hunt (resources), self (mastery) — tap fundamental human drives.
Key insights:
- Tribe = social validation; Hunt = search for resources/information; Self = personal mastery
- Predictable rewards lose power; finite variability eventually becomes predictable — aim for infinite variability
- Autonomy is critical: users must feel in control; forced engagement backfires
Product applications:
| Context | Application | Example |
|---|---|---|
| Social features (Tribe) | Variable social validation | Instagram likes — you never know how many |
| Content feeds (Hunt) | Unpredictable resource stream | Infinite scroll with algorithmically varied content |
| Gamification (Self) | Accomplishment with variable difficulty | Duolingo streaks + surprise bonus challenges |
Copy patterns:
- "See what's new" (implies variability)
- "3 people responded to your post" (tribe reward, variable quantity)
- "You've unlocked a new achievement!" (self reward, unexpected)
Ethical boundary: If users consistently feel worse after engaging (regret, time loss, anxiety), the reward system is extractive — avoid infinite scroll without natural stopping points.
See references/rewards.md when designing a reward — tribe/hunt/self patterns, the four reinforcement schedules, and reward timing. For the dopamine/anticipation mechanism behind variable rewards, see references/neuroscience-foundations.md.
4. Investment
Core concept: Users invest something — time, data, effort, social capital, money — that improves the product for next use, raises switching costs, and loads the next trigger.
Why it works: People value what they put effort into (the IKEA effect). Investment is not about immediate reward — it improves the next cycle, creating a self-reinforcing loop.
Key insights:
- Investment should come after reward, not before — users invest when they feel good
- Each investment should load the next trigger (posting content triggers reply notifications)
- Small investments compound: preferences → better recommendations → more usage; stored value grows over time
Product applications:
| Context | Application | Example |
|---|---|---|
| Data investment | History improves personalization | Spotify: more listening = better recommendations |
| Content investment | User-created content they won't abandon | Instagram posts, Notion documents |
| Reputation/social investment | Social capital that exists only on-platform | Airbnb host ratings, LinkedIn network |
Copy patterns:
- "Complete your profile to get better matches" (investment → future value)
- "The more you use it, the smarter it gets" (compound investment)
- "Invite your team to collaborate" (social investment)
Ethical boundary: Investment should genuinely improve the experience — never trap users with artificial switching costs or impossible data export; make staying the better choice through real value.
The Habit Zone
Two axes determine if a product can become a habit:
| Low Frequency | High Frequency | |
|---|---|---|
| High Perceived Value | Viable product (needs ads/marketing) | HABIT ZONE |
| Low Perceived Value | Failure | Failure |
Ask: how often do users need to engage, what's the perceived value of each engagement, and is frequency high enough to form automatic behavior?
Habit Testing
The 5% rule: a habit has formed when at least 5% of users show unprompted, habitual usage.
Three questions:
- Who are the habitual users? Which users engage most frequently, and what do they share?
- What are they doing? Identify the "Habit Path" — the action sequence that separates power users from casual users.
- Why are they doing it? What internal trigger and emotion precede usage?
See references/habit-testing.md when running the test — cohort analysis, finding the Habit Path, and confirming the 5% threshold. For worked teardowns of these loops in real products (Instagram, Slack, Duolingo, Pinterest, and failures), see references/case-studies.md.
The Manipulation Matrix
Framework for evaluating the ethics of habit-forming products:
| Maker Uses Product | Maker Doesn't Use | |
|---|---|---|
| Materially Improves User's Life | Facilitator | Peddler |
| Doesn't Improve Life | Entertainer | Dealer |
Ask: would I use this myself? Does it genuinely help users achieve their goals? Am I exploiting vulnerabilities or serving needs?
When NOT to Use the Hook Model
- Your product doesn't genuinely improve lives
- You're targeting vulnerable populations (children, addiction-prone users)
- The business model depends on user regret
- Engagement conflicts with user wellbeing
See references/ethical-boundaries.md when the Manipulation Matrix flags a concern — dark-pattern catalog and how to protect vulnerable users.
Regulatory Context
Watch emerging regulation: children's apps (COPPA, GDPR-K), dark patterns (rising FTC enforcement), "addictive" notification practices, and loot boxes (expanding gaming rules).
Onboarding Audit Checklist
Optimizing onboarding for habit formation:
First Trigger
- First action obvious and easy; right external trigger for this user
- Value proposition clear before asking for investment
First Action
- Core action completable in under 60 seconds, friction removed
- UI familiar (no new learning required)
First Reward
- Immediate feedback with a variable element (surprise, delight)
- Reward connects to an internal trigger
First Investment
- Investment asked after reward (not before), small but meaningful
- Investment loads the next trigger
Loop Completion
- Clear path back to the trigger; external triggers sent at appropriate times
- Progression through the Hook is measured
Common Mistakes
| Mistake | Why It Fails | Fix |
|---|---|---|
| Relying on external triggers indefinitely | You're renting attention, not building habits | Map product to an emotion; transition to internal triggers within 30 days |
| Making the core action too complex | Users drop off before the reward | Simplify to minimum viable action; apply the six ability factors |
| Using predictable rewards | Dopamine response fades with novelty | Add variability across tribe, hunt, and self rewards |
| Asking for investment before reward | Users haven't received value yet | Sequence: trigger, action, reward, THEN investment |
| Ignoring the ethics of your hook | User regret, backlash, regulatory risk | Use the Manipulation Matrix; be a Facilitator, not a Dealer |
Quick Diagnostic
Audit any product feature:
| Question | If No | Action |
|---|---|---|
| What's the internal trigger? | Users need reminders to use it | Research user emotions |
| Is the action dead simple? | Users start but don't complete | Remove friction |
| Is the reward variable? | Users get bored | Add unpredictability |
| Does investment load next trigger? | Users don't return | Connect investment to triggers |
Further Reading
Based on the Hook Model developed by Nir Eyal:
- "Hooked: How to Build Habit-Forming Products" by Nir Eyal
- "Indistractable: How to Control Your Attention and Choose Your Life" by Nir Eyal (companion: resisting unwanted habits and building focus)
About the Author
Nir Eyal taught at Stanford Graduate School of Business and the Hasso Plattner Institute of Design, after working in the gaming and advertising industries where he saw habit psychology firsthand. Hooked distills that research into a framework used by product teams from startups to Fortune 500s; his follow-up Indistractable addresses resisting the same triggers.
GitHub リポジトリ
よくある質問
hooked-ux Skillとは何ですか?
hooked-ux はwondelai が作成した Claude Skillです。Skillは、Claudeが必要に応じて読み込む指示とリソースをまとめ、追加の指示なしで hooked-ux に関連するタスクを実行できるようにします。
hooked-ux をインストールするには?
このページのインストールコマンドを使用してください。hooked-ux をプラグインとして Claude Code に追加するか、リポジトリを skills ディレクトリにクローンし、Claudeを再起動してSkillを読み込みます。
hooked-ux はどのカテゴリに属しますか?
hooked-ux は メタ カテゴリに属します。
hooked-ux は無料で利用できますか?
はい。hooked-ux は AIMCP に掲載されており、無料でインストールできます。
関連スキル
このスキルは、Content Collections(Markdown/MDXファイルを型安全なデータコレクションに変換するTypeScriptファーストのツール)の本番環境でテストされた設定を提供します。Zodバリデーションによる型安全性を実現し、ブログ、ドキュメントサイト、コンテンツ重視のVite + Reactアプリケーション構築時にご利用ください。Viteプラグインの設定、MDXコンパイルから、デプロイ最適化、スキーマバリデーションまで、すべてを網羅しています。
このスキルは、開発者がPolymarket予測市場プラットフォームを活用したアプリケーション構築を可能にします。API統合による取引や市場データの取得に加え、WebSocketを介したリアルタイムデータストリーミングにより、ライブ取引や市場活動を監視できます。取引戦略の実装や、ライブ市場更新を処理するツールの作成にご利用ください。
このスキルは、開発者がコマンド、ファイル、LSP操作など25種類以上のイベントタイプにフックするOpenCodeプラグインを作成することを支援します。JavaScript/TypeScriptモジュール向けに、プラグイン構造、イベントAPI仕様、および実装パターンを提供します。カスタムイベント駆動ロジックでOpenCode AIアシスタントのライフサイクルをインターセプト、監視、または拡張する必要がある場合にご利用ください。
SGLangは、高性能なLLMサービングフレームワークであり、RadixAttentionプレフィックスキャッシュを活用したJSON、正規表現、エージェントワークフロー向けの高速で構造化された生成を特長とします。特にプレフィックスが繰り返されるタスクにおいて、大幅に高速な推論を実現し、複雑な構造化出力やマルチターン対話に最適です。制約付きデコードが必要な場合や、広範なプレフィックス共有を伴うアプリケーションを構築する場合は、vLLMなどの代替案ではなくSGLangを選択してください。
