@exekutive Hello! I think that it is depend on tag name. Name of archive is <name>-<tag_name>. In your example, https://github.com/jubalh/nudoku has the latest release with tag named 5.0.0, so archive name is nudoku-5.0.0. And https://github.com/doctorfree/ninvaders has latest release with tag v1.0.0r1 that results to archive name ninvaders-1.0.0r1.

I tested this against a couple of public repos with gh release view --json tagName,name plus gh release download --archive=zip, and the generated source archive name tracked the repository name plus the release tag, for example:

• jubalh/nudoku with tag 7.0.0 downloaded as nudoku-7.0.0.zip
• doctorfree/ninvaders with tag v1.0.0r1 downloaded as ninvaders-1.0.0r1.zip

So the practical convention is close to:

<repo-name>-<tag-derived-name>.zip

But if your script needs something reliable, I still would not build too much logic around the extracted directory name.

The safer pattern is:

1. query the tag first:

gh release view --repo OWNER/REPO --json tagName

2. download to a filename you choose:

gh release download --repo OWNER/REPO <tag> --archive=zip --output source.zip

That way your script does not have to guess the archive filename or the unpacked folder name.