Who builds this
A cybersecurity student. Building in the open.
I am a cybersecurity student fascinated with AI coding harnesses, their capabilities, how we can abuse their speed and subagents, and how we can make the world a safer place to do business in an ever-growing, rapidly evolving digital age.
I build things. MoltOS, StandoutLocal, Hatchly, and PrairieCall are all mine, shipped under my own name, paid for by my own time.
The honesty beat
We remember reading a horror story on X or Reddit. An agent had left keys exposed in its logs. The logs had been committed over 50 times, and the key in them was a service-level role key. The key got revoked, but the data had been touched and the audit trail was already there. We read that thread and we thought: this is going to happen to a lot of people in the next 12 months. AI agents are fast, the developer who set up the logging is rarely the developer who notices the leak, and "I will check it before I push" is a story the industry tells itself. The lesson is not "be more careful." The lesson is "you need a smoke test that runs every time you commit." That thread is part of why RedFang exists.
Why this, why now
Hobby, interest, and a feeling of being called to use the tools I have and the knowledge I am building to make the digital world safer. Not because it is a market (though it is). Not because it is a business (though it is). Because the people shipping AI agents today are mostly first-time founders who did not know they needed a red-team, and the bad guys do.
The team
Just me. Solo. Not opposed to growing when the right person shows up. I am not the smartest person in the AI security space. I am the person who is going to make the indie version of this work, and learn out loud as I do.
If you have shipped an AI agent, a vibecoded app, or a workflow with an LLM inside, I would genuinely like to know what you found when you scanned it. The contact form is on the home page. I read every reply.