A Visual Hive & Low Code Foundation Guide
AI-Assisted Software Engineering
A Methodology That Actually Works
Start with a conversation. Document thoroughly. Execute systematically. Ship production-ready code — without the expensive mess.
19
chapters
6
parts
Free
open source
Real
projects, not theory
The Process
Five steps. Same order every time. Skip one and costs balloon. Follow them and you get production-ready code, provable outcomes, and a codebase you can actually hand to someone else.
01
Brainstorm with Opus
Real conversation in a Claude Project. Discuss scope, tech stack, architecture. Multiple turns, not one prompt.
02
Generate foundation docs
README, Architecture, rules, sprint plan, task specs. Thorough enough that AI can execute autonomously.
03
Execute in focused tasks
One task per conversation. Plan mode first. Test after. Score confidence. Close. Next.
04
Know when to start fresh
Side tasks, circular debugging, long conversations — write a task doc and start a new conversation.
05
Audit between phases
Fresh AI reviews your code. Fix what it finds before continuing.
What's in the Guide
Six parts, sequentially ordered. Read front to back or jump to what you need.
Part I
Foundation
The philosophy that makes AI-assisted development work. How to think about AI as a collaborator, and why documentation is the real product.
- Philosophy & Approach
- Tool Selection
Part II
Pre-Development
Start every project with a proper brainstorming conversation. Then generate the foundation documents that give AI everything it needs to execute.
- The Brainstorming Session
- Documentation Architecture
Part III
Execution
One task per conversation. Plan mode first. Verify and score. The repeatable workflow that produces consistent quality without hand-holding.
- The Execution Workflow
- Task Documentation
- Confidence Scoring
Part IV
Quality
Phase audits with fresh AI eyes catch what you've stopped seeing. Heavy commenting keeps future AI (and future you) in context.
- Phase Audits
- Commenting Philosophy
Part V
Advanced
Context window management, common failure patterns and how to recover from them, and how to scale the methodology to a team.
- Context Management
- Common Pitfalls
- Team Workflows
Part VI
Resources
Drop-in project templates, a tested prompt library for every phase, real case studies with numbers, and a 15-minute setup guide.
- Project Templates
- Prompts
- Case Studies
- Setup Guide
Who This Is For
Developers
Who want AI to accelerate their work without producing garbage code.
Technical Founders
Validating ideas before burning runway on the wrong thing.
Team Leads
Establishing standards for how your team uses AI tools.
Non-coders with technical sense
You can read code and guide AI even if you don't write it fluently.
Real Example
Built with this methodology
The VH Conference Toolkit — a suite of open-source tools for event professionals — was built using this exact methodology. Browse its repo to see thorough architecture docs, strict development rules, sprint-based task specs, and architectural decision records in action.
~$400
total token cost
4 weeks
to production
What this is
- ✓ A repeatable, structured methodology
- ✓ Built from real projects with real costs
- ✓ Honest about what AI can and can't do
- ✓ Works for solo devs and teams
- ✓ Free and open source
What this isn't
- ✕ A tutorial on Claude or ChatGPT basics
- ✕ Magic that removes human judgment
- ✕ A way to build with zero coding knowledge
- ✕ Another "vibe coding" hype piece
- ✕ Guaranteed success
Ready to build something properly?
30 minutes of planning saves days of rework. Every time.