Attack Vector/Scope of the attack
I read the article on arstechnica and a few things didn't sit right with me.
1) The example shows python as the language, with the poisoned payloads being present in the source code of the projects
2) That implies to me, the attack vector is people used to stackoverflow that blindly copy and paste code into their own projects.
2.1) These projects don't seem be published into a public registry since they have the same name.
2.2) People dissing github for allowing them to use the same name aren't really on the money because a fork is a fork, and a fork inherits the same name and your "user-id / org" gives it a namespaced scope in github.
3) There is nothing of course to stop the bad actors from publishing to ghcr.io under their userid/org and then doing a poison supply chain attack.
3.1) If I need to publish to maven central then I need to jump through a couple of hoops before I'm allowed to publish to the "com.myorg" group; I'm guessing that PyPi or similar don't have such restrictions
I genuinely don't know what to think of this.
Can someone help me understand precisely how this is a bigger problem than Caveat Emptor if you're just cutting and pasting shit from the web? (which is a big problem, but a _people problem not a technology problem per se_ )