Not having a binary litmus test for libertarianism doesn't mean that one thinks the term is meaningless. Just because I can't tell you what number of hairs you need to be non-bald doesn't mean I think only the absolutely hairless are bald. See http://en.wikipedia.org/wiki/Sorites_paradox.
That said, the draft St. Louis Accord arguably smuggles in both a threshold definition of libertarian and a metric for libertarian-ness. A person is a libertarian if he "wants more personal and economic liberty" and thus distinguishes himself from both the Left and the Right. A person is more libertarian to the extent that he "defends the full rights of each person to his body, labor, peaceful production, and voluntary exchanges" and shares the "ultimate goal of banishing force initiation and fraud from human relationships". I'm not saying that this threshold is deterministically binary, or that this metric imposes a total order on the set of all libertarians, because I don't think anybody can seriously claim that either is possible.
I dispute that the LP needs any more machinery for litmus-testing its representatives. Those who advocate more such machinery need to explain precisely how it would work (feel free to dry-run it on me), and in particular why it would do a better job than the Pledge and Statement of Principles have done in dissuading the LP from choosing insufficiently-libertarian representatives.
LP representatives should generally be free to disagree with any of the 27 planks of the platform, as long as they clearly advocate substantially more personal and economic liberty and don't mislead people about what policies the LP advocates. For a long list of issues on which the LP itself has taken no stand, see http://libertarianmajority.net/free-variables-in-libertarian-theory.