Zarządzanie i produkt · Planowanie
new-feature
/new-feature
Prowadzi nową funkcję od początku do wdrożenia, koordynując role zespołu od odkrycia produktu po deployment.
Dla kogo
- Autor
- andersonlimahw ↗
- Repozytorium
- lemon-ai-hub ↗
- Licencja
- MIT
- Ostatnia zmiana repozytorium
- 16.10.2026
Ocena i bezpieczeństwo
95/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść30 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
- Pliki w paczce
- 1
- Skrypty
- 0
- Komentarze HTML
- 0
- Adresy URL w SKILL.md
- 0
Przestarzałość
Brak sygnałów. Repozytorium zmienione 16.10.2026, bez odwołań do wycofanych modeli i API.
Licencja
MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.
Co robi i kiedy
Prowadzi nową funkcję od początku do wdrożenia, koordynując role zespołu od odkrycia produktu po deployment.
Oryginalny opis (EN)
End-to-end new feature workflow for the squad. Use when starting a new feature from scratch, coordinating across all squad roles from product discovery through deployment.
Jak działa
- Feature Lifecycle Overview
- Phase 1: Product Discovery (PM)
- Phase 2: UX Research & Design
- Phase 3: UI Design
- Phase 4: Architecture (Tech Lead)
- Phase 5: Backend Implementation
- Phase 6: Frontend Implementation
- Phase 7: Mobile Implementation
- Phase 8: QA
- Phase 9: DevOps & Deployment
Sekcje z nagłówków SKILL.md, w oryginale · 610 linii, 2 603 słów
Podgląd SKILL.md
---
name: new-feature
description: >
End-to-end new feature workflow for the squad. Use when starting a new
feature from scratch, coordinating across all squad roles from product
discovery through deployment.
---
# New Feature — End-to-End Squad Workflow
You are the Feature Orchestrator. You coordinate the entire squad through the complete lifecycle of a new feature, from initial product discovery through deployment and post-launch monitoring.
## Feature Lifecycle Overview
```
┌─────────────────────────────────────────────────────────────────────────┐
│ 1. DISCOVER → 2. DESIGN → 3. ARCHITECT → 4. BUILD → 5. SHIP │
│ (PM) (UX + UI) (TL) (Eng) (DevOps) │
│ │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌─────────┐ │
│ │ Problem │ │ Flow & │ │ System │ │ Backend │ │ Deploy │ │
│ │ Space │ │ Visual │ │ Design │ │ Frontend │ │ Monitor │ │
│ │ Stories │ │ Design │ │ API │ │ Mobile │ │ Rollout │ │
│ │ Criteria │ │ Specs │ │ Data │ │ QA │ │ Rollbck │ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ └─────────┘ │
│ │
│ Checkpoints: ✓ PRD Review ✓ Design Review ✓ Arch Review ✓ QA Sign │
└─────────────────────────────────────────────────────────────────────────┘
```
## Phase 1: Product Discovery (PM)
### Problem Statement
Write a clear problem statement using this template:
```markdown
## Problem Statement
**Who**: [Target user persona]
**What**: [What problem they face]
**When**: [In what context/situation]
**Impact**: [Quantified impact — lost revenue, churn, support tickets]
**Evidence**: [Data, research, or feedback supporting this]
### What we're NOT solving
- [Explicit scope boundaries]
```
### User Stories
Write user stories with clear acceptance criteria:
```markdown
## User Story: <Title>
**As a** [persona],
**I want to** [action],
**So that** [benefit].
### Acceptance Criteria
**Given** [precondition]
**When** [action]
**Then** [expected result]
**Given** [precondition — edge case]
**When** [action]
**Then** [expected result]
### Out of Scope
- [What this story explicitly does NOT cover]
### Dependencies
- [Blocked by / Blocks]
### Success Metrics
- [KPI 1]: [target] (currently [baseline])
- [KPI 2]: [target] (currently [baseline])
```
### Acceptance Criteria Checklist
Every feature MUST define:
- [ ] Happy path scenarios
- [ ] Error scenarios (invalid input, permission denied, service down)
- [ ] Edge cases (empty state, max limits, concurrent access)
- [ ] Performance criteria (response time, load capacity)
- [ ] Accessibility requirements (WCAG level)
- [ ] Analytics events to track
- [ ] Rollback criteria
---
### 🔄 Checkpoint: PRD Review
**Attendees**: PM, TL, UX Lead
**Gate criteria**:
- [ ] Problem statement is validated with data
- [ ] User stories cover happy path AND edge cases
- [ ] Acceptance criteria are testable (not vague)
- [ ] Success metrics have baselines and targets
- [ ] Dependencies are identified
- [ ] Scope is explicitly bounded
**Handoff artifact**: `PRD.md` → Shared with full squad
---
## Phase 2: UX Research & Design
### User Flow
```markdown
## User Flow: <Feature Name>
### Entry Points
1. [How users discover/access this feature]
2. [Alternative entry points]
### Flow Diagram
[Mermaid or ASCII diagram showing screens and decision points]
### Decision Points
| Step | Decision | Option A | Option B |
|------|----------|----------|----------|
| 3 | Has existing data? | Show data view | Show empty state |
### Error States
| Error | Screen | Message | Recovery Action |
|-------|--------|---------|-----------------|
| Network failure | Data load | "Unable to load" | Retry button |
### Edge Cases
- Empty state (first-time user)
- Maximum items reached
- Expired/revoked access
- Offline mode behavior
```
### Wireframes
Produce wireframes for each screen in the flow:
- Low-fidelity wireframes for layout and hierarchy
- Annotate interaction patterns (tap, swipe, long-press)
- Show responsive behavior (mobile, tablet, desktop)
- Include loading, error, and empty states
### Heuristic Review
Evaluate against Nielsen's 10 usability heuristics:
1. **Visibility of system status**: Does the user know what's happening?
2. **Match with real world**: Does it use familiar language/concepts?
3. **User control and freedom**: Can the user undo/go back?
4. **Consistency**: Is it consistent with the rest of the app?
5. **Error prevention**: Does it prevent errors before they happen?
6. **Recognition over recall**: Are options visible (not memorized)?
7. **Flexibility**: Does it serve both novice and expert users?
8. **Aesthetic and minimal design**: Is there unnecessary information?
9. **Error recovery**: Are error messages helpful and actionable?
10. **Help and documentation**: Is contextual help available?
---
### 🔄 Checkpoint: Design Review
**Attendees**: UX, UI, PM, Frontend Lead
**Gate criteria**:
- [ ] User flow covers all entry points and exit points
- [ ] All states are designed (loading, empty, error, success)
- [ ] Heuristic review completed with no critical issues
- [ ] Accessibility requirements defined
- [ ] Flow validated against acceptance criteria
**Handoff artifact**: `UX-SPEC.md` + wireframes → UI Designer, Engineers
---
## Phase 3: UI Design
### Component Specifications
For each screen, define:
```markdown
## Screen: <Name>
### Component Hierarchy
- PageContainer
- Header
- BackButton
- Title
- ActionMenu
- ContentArea
- FilterBar
- ItemList
- ItemCard (repeated)
- Thumbnail
- ItemTitle
- ItemMeta
- ActionButton
- BottomBar
- PrimaryAction
### Design Tokens Used
| Token | Value | Usage |
|-------|-------|-------|
| color.primary | #1DB954 | CTA buttons |
| spacing.md | 16px | Card padding |
| radius.lg | 12px | Card corners |
| typography.h2 | 24px/600 | Screen title |
### Responsive Breakpoints
| Breakpoint | Layout Change |
|-----------|---------------|
| < 640px (mobile) | Single column, bottom sheet actions |
| 640-1024px (tablet) | Two columns, side panel |
| > 1024px (desktop) | Three columns, inline actions |
### Interaction Specs
| Element | Interaction | Animation | Duration |
|---------|------------|-----------|----------|
| ItemCard | Tap | Scale 0.98 → 1.0 | 150ms ease-out |
| ActionButton | Tap | Ripple effect | 200ms |
| FilterBar | Scroll | Sticky at top | — |
### Visual States
| Component | Default | Hover | Active | Disabled | Loading |
|-----------|---------|-------|--------|----------|---------|
| PrimaryButton | Green bg | Darken 10% | Scale 0.97 | Grey bg, 50% opacity | Spinner |
```
### Visual Mockups
- High-fidelity mockups for each screen state
- Dark mode and light mode variants
- Motion/animation specifications
- Icon set and illustration requirements
---
## Phase 4: Architecture (Tech Lead)
### System Design
```markdown
## System Design: <Feature Name>
### Architecture Diagram
[Diagram showing services, databases, queues, and data flow]
### Components Affected
| Component | Change Type | Risk |
|-----------|------------|------|
| API Gateway | New route | Low |
| User Service | New endpoint | Medium |
| Notification Service | New event consumer | Low |
| Database | New table + migration | High |
### Non-Functional Requirements
- Latency: p99 < 200ms
- Throughput: 1000 RPS
- Availability: 99.9%
- Data consistency: Eventually consistent (acceptable delay: 5s)
```
### API Contracts
```markdown
## API: <Endpoint>
### Request
\`\`\`
POST /api/v1/playlists/{playlistId}/tracks
Authorization: Bearer <token>
Content-Type: application/json
{
"track_ids": ["string"],
"position": 0 // optional, default: append
}
\`\`\`
### Response (201 Created)
\`\`\`json
{
"playlist_id": "string",
"tracks_added": 3,
"total_tracks": 42,
"snapshot_id": "string"
}
\`\`\`
### Error Responses
| Status | Code | Message | When |
|--------|------|---------|------|
| 400 | INVALID_TRACK | "Track ID not found" | Track doesn't exist |
| 403 | NOT_OWNER | "Not playlist owner" | User lacks permission |
| 409 | DUPLICATE | "Track already in playlist" | Duplicate add |
| 429 | RATE_LIMITED | "Too many requests" | Rate limit exceeded |
### Rate Limits
- 100 requests per minute per user
- Max 100 tracks per request
### Idempotency
- Use `Idempotency-Key` header for retry safety
```
### Data Model
```markdown
## Data Model Changes
### New Tables
| Table | Purpose | Estimated Size |
|-------|---------|---------------|
| playlist_tracks | Track-playlist membership | 100M rows/year |
### Schema
\`\`\`sql
CREATE TABLE playlist_tracks (
id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
playlist_id UUID NOT NULL REFERENCES playlists(id),
track_id UUID NOT NULL REFERENCES tracks(id),
position INTEGER NOT NULL,
added_by UUID NOT NULL REFERENCES users(id),
added_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
UNIQUE(playlist_id, track_id)
);
CREATE INDEX idx_playlist_tracks_playlist ON playlist_tracks(playlist_id, position);
\`\`\`
### Migration Strategy
- Zero-downtime migration (additive only)
- Backfill strategy: N/A (new table)
- Rollback: DROP TABLE (no data loss for new feature)
```
---
### 🔄 Checkpoint: Architecture Review
**Attendees**: TL, Backend Lead, Frontend Lead, Data Engineer
**Gate criteria**:
- [ ] API contracts are complete with error cases
- [ ] Data model supports all acceptance criteria
- [ ] Non-functional requirements are defined and achievable
- [ ] Migration strategy is zero-downtime
- [ ] No single points of failure
- [ ] Monitoring and alerting plan exists
**Handoff artifact**: `ARCHITECTURE.md` + API specs → Engineering team
---
## Phase 5: Backend Implementation
### Implementation Checklist
- [ ] Database migration created and tested (up + down)
- [ ] API endpoints implemented per contract
- [ ] Business logic layer with validation
- [ ] Service layer with external integrations
- [ ] Repository/data access layer
- [ ] Input validation and sanitization
- [ ] Error handling with proper HTTP status codes
- [ ] Logging with correlation IDs
- [ ] Rate limiting configured
- [ ] Feature flag wrapping new behavior
### Test Requirements
- [ ] Unit tests for business logic (>= 80% coverage)
- [ ] Integration tests for API endpoints
- [ ] Database migration tests (up and down)
- [ ] Contract tests for external service calls
- [ ] Load tests for performance requirements
---
## Phase 6: Frontend Implementation
### Implementation Checklist
- [ ] Components built per UI spec
- [ ] State management integrated
- [ ] API integration with error handling
- [ ] Loading, error, and empty states
- [ ] Responsive design per breakpoints
- [ ] Accessibility (keyboard nav, screen reader, ARIA)
- [ ] Animations per interaction spec
- [ ] Feature flag integration
- [ ] Analytics events firing correctly
### Test Requirements
- [ ] Component unit tests
- [ ] Integration tests (user flow)
- [ ] Visual regression tests
- [ ] Accessibility automated tests (axe-core)
- [ ] Cross-browser testing
---
## Phase 7: Mobile Implementation
### Implementation Checklist
- [ ] Screens implemented per UI spec
- [ ] Navigation integrated (deep links, back handling)
- [ ] Platform-specific adaptations (iOS/Android)
- [ ] Native modules (if required)
- [ ] Offline support (if required)
- [ ] Push notification integration (if required)
- [ ] Feature flag integration
- [ ] Analytics events
### Test Requirements
- [ ] Unit tests for business logic
- [ ] UI tests for critical flows
- [ ] Device-specific testing (screen sizes, OS versions)
- [ ] Performance profiling (startup, memory, battery)
---
## Phase 8: QA
### Test Plan
```markdown
## Test Plan: <Feature Name>
### Scope
- [What's tested]
- [What's NOT tested (and why)]
### Test Matrix
| Scenario | Web | iOS | Android | API |
|----------|-----|-----|---------|-----|
| Happy path | ✓ | ✓ | ✓ | ✓ |
| Empty state | ✓ | ✓ | ✓ | — |
| Error recovery | ✓ | ✓ | ✓ | ✓ |
| Edge: max items | ✓ | ✓ | ✓ | ✓ |
| Edge: concurrent | — | — | — | ✓ |
| Performance | ✓ | ✓ | ✓ | ✓ |
| Accessibility | ✓ | ✓ | ✓ | — |
### Automation Coverage
- Automated: [list of automated test suites]
- Manual: [list of manual test cases with justification]
### Regression Risk
| Area | Risk | Mitigation |
|------|------|------------|
| Existing playlists | Medium | Regression suite |
| Search performance | Low | Load tests |
```
### QA Sign-off Criteria
- [ ] All test cases pass
- [ ] No Critical or Major bugs open
- [ ] Performance meets NFRs
- [ ] Accessibility audit passes
- [ ] Cross-platform parity verified
---
### 🔄 Checkpoint: QA Sign-off
**Attendees**: QA Lead, PM, TL
**Gate criteria**:
- [ ] Test plan executed completely
- [ ] All critical/major bugs resolved
- [ ] Performance benchmarks met
- [ ] Accessibility audit passed
- [ ] Regression suite green
**Handoff artifact**: `QA-REPORT.md` → DevOps for deployment
---
## Phase 9: DevOps & Deployment
### Deployment Pipeline
```
Code Merge → Build → Unit Tests → Integration Tests → Staging Deploy
→ Smoke Tests → Production Deploy (Canary) → Monitoring → Full Rollout
```
### Deployment Checklist
- [ ] Feature flag configured (default: OFF)
- [ ] Database migration applied to staging
- [ ] Staging smoke tests pass
- [ ] Monitoring dashboards created
- [ ] Alerts configured (error rate, latency, throughput)
- [ ] Runbook created for on-call
- [ ] Rollback procedure tested
### Monitoring Setup
```yaml
alerts:
- name: feature_error_rate
condition: error_rate > 1%
window: 5 minutes
severity: critical
action: page on-call
- name: feature_latency
condition: p99_latency > 500ms
window: 5 minutes
severity: warning
action: slack notification
- name: feature_throughput
condition: rps < 50% baseline
window: 10 minutes
severity: critical
action: page on-call + auto-rollback
```
---
## Phase 10: Data & Analytics
### Event Tracking
Define events for the feature (following event taxonomy):
```markdown
| Event | Trigger | Key Properties |
|-------|---------|---------------|
| feature_viewed | Screen loaded | source, variant |
| feature_action_taken | CTA clicked | action_type, item_id |
| feature_completed | Flow finished | duration_ms, success |
| feature_error | Error shown | error_code, step |
```
### Dashboard
Create a feature-specific dashboard:
- Adoption: DAU using feature, % of total DAU
- Engagement: actions per session, time spent
- Conversion: funnel completion rate
- Quality: error rate, latency
- Business: impact on target metric (retention, revenue)
---
## Phase 11: Launch
### Feature Flag Rollout Plan
```
Day 0: Internal dogfooding (employees only)
Day 1: 1% canary (random users)
Day 2: 5% rollout (if metrics healthy)
Day 3: 25% rollout
Day 5: 50% rollout
Day 7: 100% rollout (if no issues)
Day 14: Remove feature flag, clean up old code
```
### Launch Monitoring
| Metric | Baseline | Target | Alert Threshold |
|--------|----------|--------|-----------------|
| Error rate | 0.1% | < 0.5% | > 1% |
| p99 latency | 150ms | < 300ms | > 500ms |
| Adoption | 0% | 20% DAU | < 5% after 7 days |
| Crash rate | 0.3% | < 0.5% | > 0.8% |
### Rollback Plan
```markdown
### Rollback Triggers
- Error rate > 2% for 5 minutes
- p99 latency > 1s for 10 minutes
- Crash rate increase > 0.5%
- Data integrity issue detected
### Rollback Steps
1. Disable feature flag (instant, no deploy needed)
2. If DB migration was destructive: execute down migration
3. Notify #incidents channel
4. Create post-mortem ticket
### Rollback Owner
- Primary: [On-call engineer]
- Escalation: [Tech Lead]
```
---
## Coordination & Handoff Summary
| From | To | Artifact | Format |
|------|----|----------|--------|
| PM | Squad | PRD | `PRD.md` |
| UX | UI + Engineers | User Flows | `UX-SPEC.md` + wireframes |
| UI | Engineers | Visual Specs | `UI-SPEC.md` + mockups |
| TL | Engineers | Architecture | `ARCHITECTURE.md` + API specs |
| Engineers | QA | Build | Feature branch + test notes |
| QA | DevOps | Sign-off | `QA-REPORT.md` |
| DevOps | PM | Deployment | Rollout plan + dashboards |
| Data | PM | Analytics | Dashboard link + event spec |
Wersje
| Repozytorium | Wynik | Licencja |
|---|---|---|
| Ta strona · andersonlimahw/lemon-ai-hub | 95 | MIT |
| SteveGJones/ai-first-sdlc-practices | 100 | MIT |
Migawka bazy z 28.10.2026 · wersja paczki 8838d96094ea (skrót SHA-256 całej paczki)
Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.
Podobne skille
- 90spotify-squad
Kieruje zadania do właściwych podskilli wtyczki spotify-squad: przegląd, rytuały zespołu, backend, dane, UI i strategię QA. - 100squad-coordination
Układa wzorce koordynacji zespołu wieloagentowego, czyli podział funkcji na zadania, standupy, planowanie i zależności. - 100cross-squad
Koordynuje pracę wielu instancji Squad: wykrywanie, delegowanie i rozróżnianie, kiedy użytkownik mówi o produkcie Squad, a kiedy o grupie agentów w zwykłym znaczeniu. - 90backend-development-feature-development
Koordynuje dostarczanie funkcji backendu od wymagań po wdrożenie, gdy pracuje kilka zespołów i usług. - 91product-discovery
Przeprowadza discovery produktu od początku do końca i tworzy ustrukturyzowany PRD, także przy walidacji pomysłu lub kształtowaniu produktu. - 90using-squad-agents-skills
Wyjaśnia, od czego zacząć nowe zadanie i jak wybierać kolejne skille w cyklu pracy zespołu agentów.