The whole story around muzzle is extremely confusing, but one encounters it almost everywhere in this repo. It has 0 documentation and huge number of moving parts:

• Muzzle plugin in buildSrc folder
• Two muzzle plugins in auto-tooling module
• Muzzle task and its configuration
• Muzzle task's logs that one has to read and understand when muzzle task fails
• What does it mean "instrumentation was muzzled"??
• What is muzzle's relationship with different class loaders? I sense there is one, but I am not sure.
• What is MuzzleMatcher and how is it connected to muzzle task and muzzle plugins?
• What does the javadoc on io.opentelemetry.auto.tooling.Instrumenter.Default#getInstrumentationMuzzle actually mean? And why do we generate our own code with "compile-time bytecode transformations"?
• Muzzle plugin does some extremely convoluted things with Gradle. Such as arbitrarily reaching to other projects. This makes it impossible to use Gradle's configuration on demand. That could significantly increase build time in some cases.
• Why MuzzleVisitor uses ASM and not ByteBuddy?