imalexnet
v0.1.4
Published
I Am AlexNet - Autonomous product builder that orchestrates AI models through a 16-phase SDLC pipeline
Downloads
51
Maintainers
Readme
I Am AlexNet
Agentic CLI that builds complete products and marketing foundation by orchestrating Claude Code and Codex.
Website: imalexnet.com
Installation
npm install -g imalexnetThat's it! Now you can run alexnet from anywhere.
Development Installation
If you want to contribute or modify the code:
# Clone the repository
git clone https://github.com/yungookim/imalexnet.git
cd imalexnet
# Install dependencies
bun install
# Build and link locally (uses alexnet-dev command)
bun run dev:linkPrerequisites
Before using I Am AlexNet, you need:
Claude Code CLI
npm install -g @anthropic-ai/claude-code claude auth loginOpenAI Codex CLI
npm install -g @openai/codex codex auth login
Note: I Am AlexNet will automatically install required skills for both Claude Code and Codex during runtime. See Automatic Skill Installation for details.
Usage
# Start a new project (interactive wizard)
alexnet
# Resume an interrupted project
alexnet --resume
# Check project status
alexnet --status
# Use medium models (balanced cost/quality)
alexnet --model-tier medium
# Use light models (fastest, most economical)
alexnet --model-tier light
# Specify tech stack
alexnet --stack "nextjs,supabase,stripe"
# Disable git integration
alexnet --no-gitOptions
| Option | Description |
|--------|-------------|
| -r, --resume | Resume from last checkpoint |
| -s, --status | Show current status and exit |
| -v, --verbose | Detailed logging |
| -m, --monitor | tmux dashboard mode |
| --model-tier <tier> | "light", "medium", or "heavy" (default: heavy) |
| --stack <stack> | Tech stack override |
| --deploy <platform> | Auto-deploy (vercel, netlify) |
| --no-git | Disable git integration |
The 16-Phase Pipeline
Product Development (Phases 1-8)
- Planning - User stories and roadmap (Codex → Claude review)
- Analysis - Technical requirements (Codex → Claude review)
- Design - Architecture and prototype (Claude → Codex review)
- Implementation - Full source code (Claude → Codex review)
- Testing - Comprehensive test suite ⚠️ Perfection required
- Code Review - Security audit ⚠️ Perfection required
- Maintenance - Documentation
- Steve Jobs Review 🍎 - Product excellence check
Go-To-Market (Phases 9-14)
- Landing Page - High-converting HTML/CSS/JS
- SEO Optimization - 10+ pages, meta tags, sitemap
- Demo Video - Remotion-based explainer
- ICP & Messaging - Target personas and copy
- Google Ads - 100 ad variations CSV
- CRO Review 📈 - Conversion optimization
Finalization (Phases 15-16)
- Final Orchestrator Review 🎯 - Quality gate
- Final Dashboard 🏁 - Completion summary
Output Structure
./
├── .imalexnet/ # State and logs
├── project/ # SDLC artifacts
│ ├── planning/
│ ├── analysis/
│ ├── design/
│ ├── tests/
│ ├── reviews/
│ ├── docs/
│ └── dashboard/
├── src/ # Generated source code
├── public/ # Landing page + SEO pages
└── gtm/ # GTM artifacts
├── seo/
├── video/
├── messaging/
├── ads/
└── reviews/Cross-Model Review
Every phase (except Testing and Code Review) uses cross-model review:
- If Claude executes → Codex reviews
- If Codex executes → Claude reviews
- Maximum 5 iterations per phase
- Testing and Code Review iterate until perfect
Model Tiers
| Tier | Coding Model | Review Model | Best For | |------|--------------|--------------|----------| | heavy (default) | Claude Opus | GPT-5.2-Codex | Production apps, complex projects | | medium | Claude Sonnet | GPT-5.1-Codex-mini | Standard projects, balanced cost | | light | Claude Haiku | GPT-5.1-Codex-mini | Quick prototypes, simple apps, budget-conscious |
Automatic Skill Installation
I Am AlexNet automatically installs and configures skills for both Claude Code and Codex during runtime. This happens transparently during project initialization.
What gets installed:
- ~40 specialized skills for each model
- Skills are persisted in
.claude/settings.jsonand.codex/settings.json - Subsequent runs reuse installed skills (no reinstallation needed)
Why skills are needed: Skills enable the AI models to follow structured workflows for planning, testing, code review, and other phases. They ensure consistent, high-quality output across the 16-phase pipeline.
Installed skills:
| Skill | Claude Code | Codex | Reason | |-------|-------------|-------|--------| | commit-push-pr | ✓ | ✓ | Both commit code | | remove-ai-code-slop | ✓ | - | Claude writes implementation | | frontend-design | ✓ | - | Design + Landing page | | verification-before-completion | ✓ | ✓ | Quality gates | | test-driven-development | - | ✓ | Testing phase | | systematic-debugging | ✓ | ✓ | Bug fixing | | remotion-best-practices | ✓ | - | Video phase only | | page-cro, form-cro, etc. | ✓ | - | GTM phases | | programmatic-seo | ✓ | - | SEO phase | | paid-ads | ✓ | - | Ads phase |
First run note: The initial project may take slightly longer as skills are installed. This is a one-time setup per environment.
Development
# Run in development mode
bun run dev
# Type check
bun run typecheck
# Build
bun run build
# Build and link locally (creates alexnet-dev command)
bun run dev:linkLicense
MIT
