Ultra Builder Pro 4.4

Production-Grade AI-Powered Development System for Claude Code

---

---

Quick Start

# Clone the repository
git clone https://github.com/bivex/ultra-builder-pro.git
cd ultra-builder-pro

# Copy to Claude Code config directory
cp -r ./* ~/.claude/

# Start Claude Code
claude

---

Core Philosophy

Priority Stack (CLAUDE.md)

Priority	Rule
1	Role + Safety: Deployable code, KISS/YAGNI, think in English, respond in English
2	Context Blocks: Honor XML blocks exactly as written
3	Evidence-First: External facts require verification (Context7/Exa MCP)
4	Honesty & Challenge: Challenge user assumptions, name logical gaps
5	Architecture: Critical state must be persistable/recoverable/observable
6	Code Quality: No TODO/FIXME/placeholder, modular, avoid deep nesting
7	Testing: No mocking core logic, external deps allow test doubles
8	Action Bias: Default to progress, high-risk must brake and ask

Production Absolutism

"There is no test code. There is no demo. There is no MVP. Every line is production code. Every test is production verification."

Code Quality = Real Implementation × Real Tests × Real Dependencies
If ANY component is fake/mocked/simulated → Quality = 0

---

Workflow

/ultra-init → /ultra-research → /ultra-plan → /ultra-dev → /ultra-test → /ultra-deliver
     ↓              ↓                ↓              ↓             ↓             ↓
  Project       4-Round          Task         TDD Cycle      Quality       Release
  Setup        Discovery       Breakdown      RED→GREEN      Audit        & Deploy

---

Commands (8)

Command	Purpose	Key Features
/ultra-init	Initialize project	Auto-detect type/stack, copy templates, git setup
/ultra-research	Interactive discovery	4 rounds (User→Feature→Architecture→Quality), 90%+ confidence
/ultra-plan	Task planning	Dependency analysis, complexity assessment, context files
/ultra-dev	TDD development	RED→GREEN→REFACTOR, Codex review, auto git flow
/ultra-test	Quality audit	Anti-Pattern detection, Coverage gaps, E2E, Performance, Security
/ultra-deliver	Release preparation	CHANGELOG, build, version bump, tag, push
/ultra-status	Progress monitoring	Real-time stats, risk analysis, recommendations
/ultra-think	Deep analysis	Structured reasoning, multi-dimension comparison

Command Details

/ultra-init

• Auto-detects project type (web/api/cli/fullstack)
• Auto-detects tech stack from dependencies
• Creates .ultra/ structure with specs and tasks
• Interactive confirmation for existing projects

/ultra-research

• Round 1: User & Scenario (Personas, User Scenarios)
• Round 2: Feature Definition (User Stories, Features, Metrics)
• Round 3: Architecture Design (arc42 §1-6)
• Round 4: Quality & Deployment (arc42 §7-12)
• Each round: 6-step cycle with satisfaction rating (≥4 stars to continue)

/ultra-dev

• TDD workflow: RED (failing tests) → GREEN (pass) → REFACTOR
• Git branch management with decision tree
• Mandatory Codex review before commit
• Dual-write mode: update specs when implementation reveals gaps

/ultra-test

• Anti-Pattern Detection: Tautology, empty tests, core logic mocks
• Coverage Gap Analysis: Find untested exported functions
• E2E Testing: Chrome MCP for web UI
• Performance: Core Web Vitals (LCP <2.5s, INP <200ms, CLS <0.1)
• Security: Dependency vulnerability scan
• Auto-fix loop (max 5 attempts)

---

Skills (6)

Skill	Purpose	Modes
codex	OpenAI Codex CLI	Code analysis, refactoring, can modify code
gemini	Google Gemini CLI	Research, validation, docs; -y for code changes
frontend	Frontend development	React/Vue/Next.js, Core Web Vitals, a11y
backend	Backend development	Node.js/Python/Go, API design, security
smart-contract	Smart contract dev	EVM/Solana, security audit, Foundry tests
skill-creator	Create new skills	Workflow guidance, packaging

Codex Skill

Templates:

• research-review: Validate research output (read-only)
• code-review: Review code diff (read-only, high effort)
• test-review: Audit test suite (workspace-write)

Config: Model default, Effort medium, Sandbox workspace-write

Gemini Skill

Templates:

• tech-research: Deep research with evidence
• architecture-review: Validate architecture decisions
• documentation-gen: Generate/review documentation
• spec-validation: Validate implementation vs spec
• code-review: Review code (read-only by default)

Config: Model gemini-3-flash-preview, Mode suggest (read-only default)

Note: Use -y flag for auto-approve when code changes needed

Domain Skills

Skill	Resources
frontend	React/Vue/Next.js patterns, performance scripts, component templates
backend	API design, Node/Python/Go patterns, security audit scripts
smart-contract	ERC20/721/1155/esToken templates, vulnerability checklist, Foundry patterns

---

Quality Standards

Pre-Delivery Quality Gates

Gate	Requirement
Anti-Pattern	No tautology, empty tests, core logic mocks
Coverage Gaps	No HIGH priority untested functions
E2E	All critical flows pass
Performance	Core Web Vitals pass (frontend)
Security	No critical/high vulnerabilities
Codex Review	No critical issues

Code Limits

Metric	Limit
Function lines	≤ 50
Nesting depth	≤ 3
Cyclomatic complexity	≤ 10

Testing Policy

Type	Mock Allowed?
Core Logic (domain/service/state)	NO
Repository interfaces	NO (use testcontainers)
External APIs	Yes (testcontainers/sandbox/stub)
Third-party services	Yes (with rationale)

---

Project Structure

~/.claude/
├── CLAUDE.md                 # Main configuration (Priority Stack)
├── README.md                 # This file
├── settings.json             # Claude Code settings
│
├── commands/                 # /ultra-* commands (8)
│   ├── ultra-init.md
│   ├── ultra-research.md
│   ├── ultra-plan.md
│   ├── ultra-dev.md
│   ├── ultra-test.md
│   ├── ultra-deliver.md
│   ├── ultra-status.md
│   └── ultra-think.md
│
├── skills/                   # Domain skills (6)
│   ├── codex/                # OpenAI Codex CLI
│   ├── gemini/               # Google Gemini CLI
│   ├── frontend/             # React/Vue/Next.js
│   ├── backend/              # Node.js/Python/Go
│   ├── smart-contract/       # EVM/Solana
│   └── skill-creator/        # Create new skills
│
└── .ultra-template/          # Project initialization templates
    ├── specs/
    │   ├── product.md        # Product specification template
    │   └── architecture.md   # arc42 architecture template
    ├── tasks/
    │   ├── tasks.json        # Task registry
    │   └── contexts/         # Task context files
    └── docs/
        └── research/         # Research reports

---

Version History

v4.4.0 (2026-01-01) - Streamlined Edition

Core Changes:

• Unified Priority Stack in CLAUDE.md
• Removed agents directory (empty)
• Removed hooks system
• Removed install.sh, execution_config.json
• Removed plans/, plugins/, chrome/, telemetry/

Codex Integration:

• Added codex skill with 3 review templates
• Mandatory Codex review in /ultra-dev, /ultra-test, /ultra-research

Gemini Integration:

• Added gemini skill with 5 templates
• Default read-only mode, -y for code changes

Command Refinements:

• /ultra-test: Anti-Pattern Detection (replaced TAS scoring)
• /ultra-dev: Git branch decision tree
• /ultra-research: 4-round interactive discovery with satisfaction rating

v4.3.x (2025-12)

• Production Absolutism enforcement
• ZERO MOCK policy for core logic

---

MCP Services

Ultra Builder Pro integrates with these MCP services:

Service	Purpose
Context7	Official documentation lookup
Exa	Code examples and community practices
Chrome	E2E testing and web automation

---

License

MIT

---

Ultra Builder Pro: No mock. No demo. No MVP. Production-grade only.