In contrast to digital cryptographic algorithms and protocols where many qualified high-IQ individuals dig into the details and specifics, physical tamper resistance for low-tech packages is not pentested as much.
There are some folks out there who do it for money, though, but their results aren't easily available. There's one link to a Vulnerability Assessment Team at Argonne National Lab with intriguing results: http://www.ne.anl.gov/capabilities/vat/seals/index.html
I have looked for patents at U.S.Patent and Trademark Office and identified quite a few of them, the latest being U.S.Patent 7,230,687 B2, and the earliest dating back to 1916 (U.S.Patent 1,201,519 assigned to one Arvid Sorensen).
I wonder what criteria should one apply to various tamper-resistant offers in the market (numbering upward of 180, based on a shopping engine search), provided that there are no in-house skilled physical pentesters.
One criterion that I've already hinted at is presence or absence of a patent (of course it doesn't prove security - there are loads of patented snake oil recipes, for that matter, but at least it is somewhat analogous to open-source in the digital world).
Another solution would be to sample large outfits' secure tamper-resistant envelopes, hopefully piggy-backing on their evaluations and "going with the flow", but I'm not sure they haven't done the same.
Please note that this is not a shopping question. Answers should avoid naming providers and instead present objective criteria for evaluation.