Research, Plan, Implement: Our Agent-Driven Development Cycle
Part of SDLC2: the problem with planning in the dark, the cycle we run instead, and how it looks in our backend, with the artifact chain that makes it stick.
ls posts/
All Posts
10 entries
Agents as Code: the abstraction layer has moved on, did you?
Part of SDLC2: one repo for our agent commands and skills, changes via PR, and the unexpected joy of writing them in plain English together.
SDLC2: How We're Changing How We Build Products
At Integral we're changing how we build products, software and beyond. This is where we document our experience, our simple approach, and our wins.
How to Love Lovable: Building Production-Ready Infrastructure Around AI Code Generators
How we built Jantro, a Firebase-hosted wrapper that gives Lovable apps production-grade auth, access control, and deployment without sacrificing speed.
Just Makes AI Smarter: Stop Reinventing Commands Every Time
Your command runner is a project interface. Make it readable for AI tools and watch them stop improvising your build process.
The Captain Role: Driving Features Forward Without Becoming a Bottleneck
How one person takes ownership of feature delivery without turning into a micromanager or single point of failure.
Guardian Role: Protecting Focus in a World of Interruptions
The rotating role that shields your team from interruptions so they can actually build things instead of constantly context-switching.
Weekly Cycles: Rhythm Without Rituals
How we structure weeks to balance planning, building, and shipping without daily standups or agile theater.
How We Work: Building the Engineering Team of the Future
Welcome to our engineering blog. Here's our philosophy on building small, senior teams that ship fast without bureaucracy.
Plan First: Why Human Planning Still Matters in the AI Age
Detailed planning isn't overhead when AI is writing code; it's the difference between useful output and expensive garbage.