
We went from an idea to a working public product in roughly three months. This is where I really love how they stand out, because every person in their team was self-organizing and could go from 0-100 in no time at all. That included the initial platform, the product interface and user-facing documentation. We were then able to expand the product into the Playground and MCP ecosystem rather than stopping at the initial deployment product. The hackathon produced more than 50 potential use cases within the first 2–3 weeks of launch. That was useful because it gave us real evidence of how developers were thinking about using agents and MCPs, rather than relying entirely on our assumptions.
The biggest thing was their ability to move between product and engineering. Illegal Agents was being built at a time when AI agent frameworks and MCP were still moving very quickly. There wasn't a mature set of tools or established patterns for everything we wanted to do. Grayhat was comfortable working through those problems directly. They could take a product idea, turn it into a design, and then go deep into the implementation when the technical problem required it. They also weren't afraid to work in open-source repositories and deal with unfamiliar codebases. That flexibility mattered more to us than simply having a team that could execute a predefined backlog.





















