From a practicality standpoint, writing in TS allows you to execute arbitrary JS in-process with a simple exec() call; I have no doubt they'll use this to more deeply integrate agents with existing codebases in the near future. E.g. agents rather than just reading the code, will be able to directly import the data structures themselves, use libraries within the JS/TS ecosystem to parse code into an AST, and execute in-process test harnesses to validate behavior while editing.

And the MCP field is already pretty heavily saturated with TypeScript and JSONSchema, so using TS for it is a very ergonomic experience. Also since it's written in TypeScript, it's much more easy to integrate it with editors like VSCode (or Google's new Antimatter) which are built on top of Electron.

Are they using it this way? I think if you have access to the cli you can do whatever you want, and it already does use tools and use libraries and parse code. It's just using the cli. Not to mention it's much better this way because I can see it doing all these things. I can see what commands it wants to run and it asks me about it.