Mannay Claude Code

Skills Reference

Complete reference for all Mannay skills (systematic workflows).

Overview

Skills are systematic workflows that auto-activate based on task type. They ensure best practices and quality standards.

┌─────────────────────────────────────────────────────────────────────┐
│                          USER REQUEST                                │
│                    "Add user authentication"                         │
└─────────────────────────────────┬───────────────────────────────────┘
                                  │
                                  ▼
┌─────────────────────────────────────────────────────────────────────┐
│                         USING-MANNAY                                 │
│                    (Meta-Orchestrator)                               │
│                                                                      │
│   1. Detect task type: NEW FEATURE                                  │
│   2. Activate mandatory skill: brainstorming                        │
│   3. Identify domains: auth, API, security                          │
│   4. Queue agents: security-engineer, backend-architect, api-designer│
└─────────────────────────────────┬───────────────────────────────────┘
                                  │
          ┌───────────────────────┼───────────────────────┐
          │                       │                       │
          ▼                       ▼                       ▼
┌─────────────────┐    ┌─────────────────┐    ┌─────────────────┐
│  BRAINSTORMING  │───►│  TASK-ANALYSIS  │───►│  WRITING-PLANS  │
│  Design first   │    │  Break it down  │    │  Detailed steps │
└─────────────────┘    └─────────────────┘    └─────────────────┘
                                                       │
                                                       ▼
┌─────────────────────────────────────────────────────────────────────┐
│                    IMPLEMENTATION PHASE                              │
│                                                                      │
│   ┌─────────────────────────────────────────────────────────────┐  │
│   │                         TDD CYCLE                            │  │
│   │    Write Test ──► Run (Fail) ──► Code ──► Run (Pass) ──► ♻️ │  │
│   └─────────────────────────────────────────────────────────────┘  │
│                              │                                      │
│   ┌──────────────────────────┼──────────────────────────┐          │
│   │                          │                          │          │
│   ▼                          ▼                          ▼          │
│ ┌────────────────┐  ┌────────────────┐  ┌────────────────┐        │
│ │security-engineer│  │backend-architect│  │  api-designer  │        │
│ │ Review auth    │  │ Schema design  │  │ Endpoint specs │        │
│ └────────────────┘  └────────────────┘  └────────────────┘        │
│                                                                      │
└─────────────────────────────────────────────────────────────────────┘
                                  │
                                  ▼
┌─────────────────────────────────────────────────────────────────────┐
│                          GIT COMMIT                                  │
│              feat(auth): add user authentication                     │
└─────────────────────────────────────────────────────────────────────┘

Skill Categories

┌─────────────────────────────────────────────────────────────────────────────┐
│                               SKILLS                                         │
├─────────────────┬────────────────┬────────────────┬────────────────────────┤
│   MANDATORY     │   PLANNING     │   EXECUTION    │    QUALITY/DESIGN      │
│   (Auto-On)     │                │                │                        │
├─────────────────┼────────────────┼────────────────┼────────────────────────┤
│ • TDD           │ • task-analysis│ • executing-   │ • code-review          │
│ • debugging     │ • feature-     │   plans        │ • api-testing          │
│ • brainstorming │   planning     │ • ralph-mode   │ • frontend-design      │
│ • git           │ • writing-plans│                │ • adaptive-design (*)  │
│ • adaptive-     │ • prd-builder  │                │ • ui-animation (*)     │
│   design (*)    │                │                │ • compound-eng         │
│ • ui-animation  │                │                │ • web-design-guidelines│
│   (*)           │                │                │ • react-best-practices │
└─────────────────┴────────────────┴────────────────┴────────────────────────┘
                     (*) MANDATORY for any visual work

Mandatory Skills (Auto-Activate)

These activate automatically - no trigger needed:

test-driven-development

Activates: Any code change

Process: RED-GREEN-REFACTOR

  1. Write failing test
  2. Verify it fails
  3. Write minimal code to pass
  4. Verify it passes
  5. Refactor
  6. Commit

Integration:


systematic-debugging

Activates: Any bug, error, or unexpected behavior

Process: 4-Phase Root Cause Analysis

  1. Phase 1: Root cause investigation
  2. Phase 2: Pattern analysis
  3. Phase 3: Hypothesis and testing
  4. Phase 4: Implementation with TDD

Integration:


brainstorming

Activates: New feature or significant change

Process:

  1. Understand project context
  2. Ask clarifying questions (one at a time)
  3. Propose 2-3 approaches
  4. Present design in sections
  5. Invoke relevant agents
  6. Document decision

Integration:


git

Activates: Commits, pushes, branches, PRs

Process: Conventional commits

Integration:


Planning Skills

task-analysis

Triggers: “implement this”, “build this”, “break this down”

Process:

  1. Detect project context
  2. Analyze task complexity
  3. Create implementation plan
  4. Recommend agents

Integration:


feature-planning

Triggers: “plan this feature”, “feature spec”

Process:

  1. Feature breakdown (user stories, requirements)
  2. Technical specification
  3. Implementation plan (6 phases)
  4. Invoke relevant agents

Integration:


writing-plans

Triggers: “write a plan”, “detailed plan”

Process:

  1. Analyze requirements
  2. Create step-by-step plan
  3. Define verification criteria
  4. Structure for execution

Integration:


executing-plans

Triggers: “execute this plan”, “implement the plan”

Process:

  1. Load plan
  2. Execute in batches
  3. Verify after each batch
  4. Report progress

Integration:


PRD & Ralph Skills

prd-builder

Triggers: “create a prd”, “write prd”, “spec out”

Process: 6 Phases

  1. Requirements discovery
  2. Design exploration
  3. Technical planning
  4. Story decomposition
  5. PRD generation (JSON)
  6. Quality validation

Integration:


ralph-mode

Triggers: “run ralph”, “start ralph”, “autonomous mode”

Process: 7 Phases per iteration

  1. Context loading
  2. Story selection
  3. Story execution (TDD)
  4. Verification
  5. Commit (mandatory)
  6. Update files
  7. Completion check

Integration:


Quality Skills

requesting-code-review

Triggers: “review this code”, “code review”

Process: Two-stage review

  1. Spec compliance review
  2. Code quality review

Integration:


api-testing

Triggers: “test this api”, “api tests”

Process:

  1. Happy path tests
  2. Error cases
  3. Edge cases
  4. Security tests

Integration:


Design Skills

frontend-design

Triggers: “build a page”, “create component”, “design UI”

Process:

Integration:


adaptive-design (MANDATORY for visual)

Activates: Any visual component, page, or UI element

Process:

  1. Design with real data (not idealized mocks)
  2. Test edge cases: empty, minimal, maximum content
  3. Use constraint-based layouts, not fixed pixels
  4. Handle missing images, long text, overflow
  5. Responsive across 320px - 2560px viewports

Key Principles:

Integration:


ui-animation (MANDATORY for visual)

Activates: Any visual component, page, or UI element

Philosophy: First ask “Should this animate at all?” based on interaction frequency.

Process:

  1. Check frequency (100s/day = no animation, keyboard = never)
  2. Apply precise timing (all UI < 300ms except drawers)
  3. Use enter recipe: opacity + translateY + blur
  4. Make exits subtler than enters
  5. Scale from 0.93+ (never 0)
  6. Use correct easing (90% ease-out, custom curves)
  7. Set transform-origin from trigger point
  8. Use transitions (interruptible), not keyframes
  9. Respect prefers-reduced-motion

Quick Reference:

Philosophy:       Should this animate? Check frequency first.
Button press:     scale(0.97), 150ms, ease-out
Tooltip:          scale(0.97), 125ms, ease-out
Modal:            scale(0.93), 200ms, ease-out + blur
Dropdown:         scale(0.93), 180ms, ease-out, origin from trigger
Drawer (iOS):     500ms, cubic-bezier(0.32, 0.72, 0, 1)
Toast dismiss:    4 seconds (pause when inactive)

Enter:            opacity + translateY(8) + blur(4px)
Exit:             SUBTLER (translateY -8px, blur 2px)

References:

Integration:


compound-engineering

Triggers: “systematic development”, “compound work”

Process: 4-Phase Loop

  1. Plan (40%)
  2. Work (20%)
  3. Review (20%)
  4. Compound (20%)

Each unit of work makes future work easier.


web-design-guidelines

Triggers: “ui design”, “ux review”, “design audit”, “accessibility check”

Impact: HIGH

Process:

Integration:


react-best-practices

Triggers: “React”, “Next.js”, “performance”, “re-render”, “bundle size”

Impact: CRITICAL

Process: 45 rules across 8 priority categories

  1. Eliminating Waterfalls (CRITICAL) - async patterns, Promise.all
  2. Bundle Size Optimization (CRITICAL) - dynamic imports, barrel avoidance
  3. Server-Side Performance (HIGH) - React.cache(), LRU caching
  4. Client-Side Data Fetching (MEDIUM-HIGH) - SWR deduplication
  5. Re-render Optimization (MEDIUM) - memo, state patterns
  6. Rendering Performance (MEDIUM) - content-visibility, SVG
  7. JavaScript Performance (LOW-MEDIUM) - loops, caching
  8. Advanced Patterns (LOW) - refs, callbacks

Integration:


Orchestrator

using-mannay

Purpose: Meta-orchestrator for all skills and agents

Behavior:

Integration: