Some of the tool's shell scripts need to be run as the www user and some not.
virtualenv_activate.sh checks that the current user is the correct one and if not, runs exit 1. Because we've sourced this script, exit 1 quits the session entirely, booting the user out of vagrant ssh. We shouldn't do this, but should retain the behaviour of providing the helpful error message and quitting the script(s) being run.
The task as originally written was already the case - virtualenv_activate.sh checks the user. However, when it exits it quits vagrant ssh too. Changing task scope.
So, while fixing this problem in a branch, I encountered/reproduced the originally reported issue. The scripts that directly use a virtualenv do in fact have the user check thanks to the virualenv_activate script. Some of the wrapper scripts that compose those into useful workflows don't because the test for the "right thing" needs to be a little more complicated.