A python VM that consumes bytecode might be doable in not-ludicrous-amounts of C. Not 500 lines I suppose. But something manageable I think? Especially if you targeted the older releases.
https://github.com/python/cpython/blob/main/Python/ceval.c
It's 3619 lines. It's explained in this 515 line file:
https://github.com/python/cpython/blob/main/InternalDocs/
For comparison, there is a pure Python bytecode interpreter, its VM is here:
https://github.com/nedbat/byterun/blob/master/byterun/pyvm2....
It's 1043 lines.