si14agents.com: Trying to Bring an Idea to Life

There's this project that's been sitting in my mental backlog - si14agents.com. I vibe coded it on Christmas Eve just to clarify my own thinking about the concept. What was this thing really about? Where was I trying to go with it?

January was experimental. I built my first real code agent - reverse-engineered from how Bolt works - that could connect to an API and actually solve tasks on GitHub. Messy, but it worked.

Then February and March went quiet. Lots of thinking, less building.

But April got interesting. I started building an MCP client, which evolved into an MCP Client paired with an A2A Agent Framework. I hacked together a registry of POC agents - little experiments I'd built to test concepts.

The last two weeks, I've been obsessing over aiguardrails.com (not live yet) and really drilling down on optimizing my Cline workflow with rules and ADRs. I'm hitting a point where 9 out of 10 prompts just work - green builds, passing tests, functional code.

So the goal is to bring it all together under si14agents.com as the umbrella concept. Get signup working, get payments sorted, build something that actually feels like a product instead of scattered experiments.

It's probably the most ambitious hobby project I've tackled. Ambitious enough that sometimes I question whether I'm overextending. But that's the whole point of this challenge, right? To push beyond what feels comfortable and see what's actually possible when you combine good prompting with solid architectural thinking.

The concept deserves to exist as more than sketches and experiments.