Mass assignment is one of the most prominent vulnerabilities in RESTful APIs that originates from a misconfiguration in common web frameworks. This allows attackers to exploit naming convention and automatic binding to craft malicious requests that (massively) override data supposed to be read-only.
In this paper, we adopt a black-box testing perspective to automatically detect mass assignment vulnerabilities in RESTful APIs. Indeed, execution scenarios are generated purely based on the OpenAPI specification, that lists the available operations and their message format. Clustering is used to group similar operations and reveal read-only fields, the latter are candidates for mass assignment. Then, test interaction sequences are automatically generated by instantiating abstract testing templates, with the aim of trying to use the found read-only fields to carry out a mass assignment attack. Test interactions are run, and their execution is assessed by a specific oracle, in order to reveal whether the vulnerability could be successfully exploited.
The proposed novel approach has been implemented and evaluated on a set of case studies written in different programming languages. The evaluation highlights that the approach is quite effective in detecting seeded vulnerabilities, with a remarkably high accuracy. PDF version of the paper.