Rather than installing this via npm or yarn, I believe deno should be the first-class installation method. It'd be more forward-thinking, I think, to write against deno, another "reboot" by an influential author, and it would help differentiate from the current Babel/Webpack/ESLint toolchain. I believe this should be achievable especially since you are aiming for zero dependencies, and the project is still very young.
Deno already has its own built in tools to deal with these issues. If anything, feels like an unnecessary addition considering the idea would be to make the experience simpler for the user.
I know I closed this, but I wanted to add for posterity that deno bundle isn't currently intended for browsers:
deno bundleis for running code in Deno.
https://github.com/denoland/deno/issues/6516#issuecomment-650756464
Other features like linting would still be a clash, though.
Most helpful comment
Deno already has its own built in tools to deal with these issues. If anything, feels like an unnecessary addition considering the idea would be to make the experience simpler for the user.