> Can I safely assume that immediately after the JVMTI interface
> RunAgentThread() returns to the caller, that the new thread's state
> will be something other than 0 or "NEW" (i.e. not java.lang.Thread.State.NEW)?

Yes, the thread has been started and thus will be alive unless the the
thread died immediately.

> I want to call SetThreadLocalStorage() immediately after RunAgentThread(),
> and I would rather not have to deal with JVMTI_ERROR_THREAD_NOT_ALIVE errors.
> It appears to be safe, but just thought I'd ask. I didn't see anything
> in the doc on RunAgentThread() that clarified this for me.