When decisions aren’t captured, your codebase forgets how it was built.

If you're defining a role: subagent with skills. The agent is the long-lived actor, skills are its reference material If you're running a task: skill with context: fork. Isolation without spinning up a full agent file Full breakdown on how to combine both effectively 👇

Why is everyone talking about background agents and software factories?

Companies are building 24/7 software factories powered by background agents. And the results are already showing.

From SDLC -> Software Factory Loop! 1 or 4

From SDLC -> Software Factory Loop! 2 or 4

From SDLC -> Software Factory Loop! 3 or 4

From SDLC to Software Factories: The Next Evolution of Software Development

Truth: AI is advancing at crazy speeds. Everyday we wake up to new realities. Question: How do you balance exponential growth, safety & future possibilities, all at the same time? Chatted about this and more, with Anthropic CTO Rahul Patil South Park Commons India. Full EP out!

I’d encourage teams to focus on building custom skills on top of LLMs, instead of buying dozens of agents. Here’s why: Most agents are stateless. They rely on prompts and limited context, so every task starts from scratch. Skills change that. Skills are emerging as the way to

Delightful writeup!

Why are all big tech companies building software factories?

What are the building blocks for an AI-Native Software Factory?

Companies like Ramp and Stripe are building software factories using a unified context layer.