This question already has an answer here:
I have recently had to submit a number of confidential PDF documents to a website. These documents contain more than enough information to use in identity theft, and I can conceive how others using the same site would upload even more information.
When viewing these documents back through the website, I noticed that the method of protecting the content from unauthorised access is to use a long, apparently random string:
I also note that this content is susceptible to sslstrip type attacks i.e. http can be used instead of https if forced using a MITM.
The lack of genuine access control to the document concerns me. Browser history, link sharing, sniffing if http is used; these all leak the document and how to access it.
There is already a question and response along these lines "Are random URLs a safe way to protect profile photos?", however, in this instance:
- The asset is information that could lead to identity theft, rather than a profile picture.
- A profile picture should be accessible by many others in most cases (at least anyone else signed into the same website, at most everyone). The documents in this case should at most be visible to three parties and be private from everyone else.
- There is an attempt to conceal the URL using https which can easily be subverted.
- The specific application is for independent adjudication is tenancy deposit disputes. The only parties who should see the documents are the tenant, landlord and adjudicator.
Is this considered adequate security?