This just looks like a really cryptic way to do startsWith + length. Am I correct in saying that's what your code is doing?

I mean, you could also interpret it that way, but I think it's less cryptic than startsWith + length.

To me, this version is much clearer:

function isATextFileNamed(filename, name) {
  return filename.endsWith('.txt') && filename.slice(0, -4) === name;
}

Than if you wanted to use startsWith and length:

function isATextFileNamed(filename, name) {
  return filename.endsWith('.txt') && filename.startsWith(name) && filename.length === name.length + 4;
}

Another use-case is if you want to check a first line.

This:

function firstLineEquals(text, firstLine) {
  const newlinePosition = text.indexOf('\n');
  return text.slice(0, newlinePosition) === firstLine;
}

Is much clearer than this:

function firstLineEquals(text, firstLine) {
  const newlinePosition = text.indexOf('\n');
  return text.startsWith(firstLine) && newlinePosition === firstLine.length;
}

But the former also dings prefer-string-starts-ends-with.

To me, this version is much clearer

IMO it's not because (in my experience) most devs don't have experience with negative index form of slice, which means there's non-trivial overhead in understanding that slice(0, -4) means "take all but the last 4 characters". TBH if I wasn't paying enough attention, I could easily misread this as substr(-4), which has a very different meaning.

For your second example, I'd use substring instead, as I think it communicates the intention more clearly and is easier to read than slice.

But this is just bikeshedding.
I was more just double checking I understood your code's intention.

The rule doesn't currently understand negative indices in the second position for slice.
Happy to accept a PR!