I want to release v1.0 of this module when the following conditions are met, and not before:

1. It's feature-complete, i.e.:

• a fs.FS implementation is available, and operational, for all filesystems targeted for support (see README.md)
• context can be inserted into all filesystems where applicable
• HTTP headers can be inserted into all filesystems where applicable
• Content-Type guessing works in all filesystems where applicable

2. It's stable and well-tested on all supported platforms (linux/darwin/windows, and amd64/arm64)
3. The API is good enough (i.e. not awkward to use, leaks as few abstractions as possible)