# Human Centered Feature Design

> Human-Centered Feature Design Skill

- Skill: `lillyliujun/human-centered-feature-design` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add lillyliujun/human-centered-feature-design`
- Raw SKILL.md: https://api.skillmd.com/api/skills/lillyliujun/human-centered-feature-design/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: LillyLiuJun (https://skillmd.com/u/lillyliujun)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/lillyliujun/human-centered-feature-design

---


# Human-Centered Feature Design Skill

## Foundational Worldview (The Starting Point of Design)

**Humans are not variables of products; products are variables of humans.**

Three Core Postulates:
- **Humans are self-regulating systems (Homeostasis)** — Human needs arise when an individual deviates from internal equilibrium. Products exist to help users restore balance. A product cannot manufacture need out of thin air; it can only become an outlet through which existing tension resolves.
- **Humans are cognitively fragile, yet crave a sense of agency** — The system’s responsibility is to scaffold and sustain the user while attributing all credit back to them. A truly great system remains invisible.
- **Humans always exist within situated contexts** — Whether any signal takes effect depends entirely on how closely it matches the user’s immediate situation. A product’s influence is borrowed from the context, never owned.

These three postulates govern the evaluation criteria: **Does this feature genuinely help the user restore balance? Does it leave the user feeling capable and in control? Does it surface within the right context?**

---

## Workflow & Usage

When a user describes a feature concept, execute the following diagnostic and design workflow.

### Step 1: Gather Prerequisite Context

Before diving into solutions, establish the following three vectors:
1. **What fundamental need does this feature address?** In which direction has the user deviated from balance (*Deficiency* or *Overflow*)?
2. **Is this need Discrete or Continuous?** Can its completion be objectively verified?
3. **What is the target user's typical situated context?** (Physical space, temporal moment, and baseline cognitive/emotional load).

*If the initial description is ambiguous, clarify these three questions first before proceeding.*

### Step 2: Deliver Comprehensive Specification

The specification must thoroughly address the following three architectural modules.

---

## Module 1: Motivation Design — Sustaining the Feature as the Natural Outlet for Need

### 1. Diagnose Need Topology

| Need Type | Defining Trait | Archetypal Examples | Primary Retention Mechanism |
|---|---|---|---|
| **Discrete** | Completion is objectively verifiable; once resolved, active motivation terminates immediately. | Weather lookup, translation, point-to-point navigation | **Trade Outcomes for Trust** — Delivering outcomes that exceed expectations ensures this tool is the immediate first recall next time the need arises. |
| **Continuous** | No external benchmark marks completion; motivation surfaces, recedes, and resurfaces perpetually. | Loneliness, boredom, desire for creative expression, curiosity | **Trade Continuity for Stickiness** — Satisfy the immediate urge while subtly introducing a new anticipated loop, ensuring unfinished momentum remains. |
| **Hybrid** | Grounded in objective milestones, but the ceiling can extend indefinitely. | Fitness, language learning, skill acquisition | Combine both mechanisms systematically. |

### 2. Craft Motivation Outlets by Topology

**For Discrete Features:**
- **Outcome Precision is King**: Every single execution must meet or surpass the anticipated threshold.
- **Manage the Trust Depletion Curve**: As usage count rises, user tolerance drops and expectation escalates; outcome precision must never degrade.
- **Occupy Mental First Place**: Condition the feature to become a reflexive impulse whenever the specific need flares up.

**For Continuous Features:**
- **Deliver Genuine Resolution Now**: The immediate urge must be truly answered; superficial delay will cause instant churn.
- **Introduce Subtle Disequilibrium at Balance**: Exactly as the user regains equilibrium, plant the seed of the next "unfinished momentum."
- **Keep Anticipation Fresh**: Refresh the value aperture before saturation or hedonic adaptation sets in.

**Core Litmus Test: Is this feature helping the user restore internal balance, or is it manufacturing artificial anxiety?**

---

## Module 2: Sense of Agency Design — Feeling Capable Across Every Interaction Touchpoint

A genuine sense of agency is not granted by praise; it emerges naturally when every stage of interaction is calibrated with dignity. Design across the six sequential touchpoints:

### Stage 1: Perception
**Goal: Transform vague sensing into concrete recognition.**
- Does the opening viewport immediately communicate: *"Yes, this is exactly what I needed right now"*?
- Do the visual language, copy, and atmospheric weight resonate with the user’s ambient cognitive state?
- Does the entry point point directly toward the user’s intended direction of relief?

### Stage 2: Search & Association
**Goal: Align the user's mental search path with the system’s topology.**
- Does the information architecture conform to pre-existing mental models rather than internal database schemas?
- Do visual hierarchies minimize cognitive energy expenditure?
- Is the path for newcomers intuitively self-evident without explicit tutorials?

### Stage 3: Decision
**Goal: Define conscious boundaries — minimize friction by default; provide absolute safety when stakes are real.**

*Three conditions requiring deliberate "System 2" cognitive engagement:*
- High perceived cost (money, time investment, sensitive personal data).
- Outcome uncertainty (user cannot predict what happens upon interaction).
- Irreversibility (user fears the action cannot be cleanly undone).

- Are there unnecessary decision forks along the primary Golden Path?
- Do high-cost, uncertain, or irreversible actions provide explicit reassurance, previews, and fallbacks?

### Stage 4: Action
**Goal: Instant physical and cognitive feedback — prevent continuity fractures.**
- Does every tap or input receive immediate, unambiguous state acknowledgement?
- Are loading states contextualized? Do error states provide actionable, non-blaming recovery paths?

### Stage 5: Result Evaluation
**Goal: Precise promise upon entry; beyond-expected delivery upon exit.**
- **Upstream Alignment**: Did signals during the Perception stage strictly align with what the feature actually delivers?
- **Downstream Resonance**: Is the delivered outcome compelling enough to foster deeper trust and evoke desire for more?
- **Attribution**: Does the presentation celebrate *"I did this!"* rather than flaunting *"Look what the clever algorithm did for you"*?

### Stage 6: Habit Formation
**Goal: Automate the Golden Path; prepare secondary depth pathways before adaptation.**

**The Golden Path (Habitualization):**
- Trigger + Repetition + Meaningful Reward — keep the path minimal and frictionless; ensure consistent positive outcomes.
- Apply only subtractions and polish to the Golden Path; avoid disruptive redesigns.

**Overlay Pathways (Deepening Relationship):**
- Unveil adjacent value propositions precisely when user satisfaction is at its peak.
- Low cognitive barrier (covered by existing mental model) + Clear stimulus (contextually relevant signal).

---

## Module 3: Trigger Design — Activating the User Within the Right Context

### Step 1: Deconstruct Situated Context

| Dimension | Core Definition | Design Implication |
|---|---|---|
| **Space** | Where is the user physically situated? | Dictates available attention bandwidth and physical interaction constraints. |
| **Time** | What temporal window is this? | Dictates baseline biological rhythms and psychological reserves. |
| **Psychological State** | Ambient mood and current cognitive load. | Dictates what frequency of signal can be processed without becoming hostile noise. |

### Step 2: Identify Gaps via the Fogg Behavior Model (FBM)

Within the user’s situated context, diagnose which prerequisite is deficient:
- **Motivation**: Is the tension or appetite strong enough right now?
- **Ability (Sense of Agency)**: Does the user perceive that they have the capability and friction-free bandwidth to execute right now?
- **Prompt / Timing**: Is this specific moment the natural window for a trigger?

### Step 3: Engineer Contextual Scaffolding

Variables within direct product agency: Information density, atmospheric mood, operational pacing, social proof, and notification dispatch.

- **Deficiency in Motivation**: Amplify relevant resonance (contextual framing, social resonance, unfinished progress reminders).
- **Deficiency in Ability**: Eliminate uncertainty (set crisp time/outcome expectations, transparent guidance, reversible previews).
- **Deficiency in Timing**: Harmonize with external rhythms (intelligent notification timing, contextual ambient home states).

**Continuous Activation**: Contexts evolve; behavioral bottlenecks shift. Never spam repetitive static signals. Continuously diagnose new situational gaps and deploy attuned nudges.

---

## Deliverable Format

```markdown
## 1. Feature Positioning & Taxonomy
[Need Type Diagnosis + Direction of Disequilibrium + Core Reason for Feature Existence]

## 2. Motivation Architecture
[Need Topology → Retention Loop Mechanics → Explicit Design Recommendations]

## 3. Sense of Agency Architecture
[Critical Decisions Across the 6 Touchpoints, Highlighting High-Risk Stages]

## 4. Trigger & Contextual Scaffolding
[Situated Context Breakdown → FBM Gap Diagnosis → Concrete Trigger Mechanism]

## 5. Antifragile Risk Assessment
[Decisions Most Likely to Erode Continuous Engagement & Concrete Mitigations]
```

---

## Evaluation Benchmark

Evaluate every product decision against these three immutable questions:

1. **Does this decision help the user restore balance, or does it engineer artificial anxiety?**
2. **Does this decision make users feel sovereign and in control, or dragged around by system mechanics?**
3. **Is this decision embedded naturally within the user’s real-world context, or is it generating intrusive noise?**

Only when all three answers favor the former is the design validated as human-centered.