Cross Site History Manipulation (XSHM)
Contributor(s): James Bowie, kingthorin
Description
Cross-Site History Manipulation (XSHM) is a SOP (Same Origin Policy) security breach. SOP is the most important security concept of modern browsers. SOP means that web pages from different origins by design cannot communicate with each other. Cross-Site History Manipulation breach is based on the fact that client-side browser history object is not properly partitioned on a per-site basis. Manipulating browser history may lead to SOP compromising, allow bi-directional CSRF and other exploitations such as: user privacy violation, login status detection, resources mapping, sensitive information inferring, users’ activity tracking and URL parameter stealing.
Risk Factors
By manipulating the browser history it is possible to compromise SOP and violate user privacy. Using CSRF in conjunction with history manipulation, not only integrity but also confidentiality can be targeted. Feedbacks from a different origin can be accessed and Cross-Site information leakage is achieved.
The following attack vectors based on techniques of XSHM are possible:
- Cross-Site Condition Leakage
- Login Detection
- Resource Mapping
- Error Leakage
- State Detection
- Information Inference
- Cross-Site User Tracking
- Cross-Site URL/Parameters Enumeration
Examples
What is Condition Leakage?
Condition leakage occurs when an attacker can infer a sensitive value of a conditional statement in an attacked application. For example, if a site contains the following logic:
Page A: If (CONDITION)
Redirect(Page B)
an attacker can execute the CSRF and get an indication about the value of the condition as a feedback. This attack is executed from an attacker site. The site then submits a Cross-Site request to a victim site, and by manipulating the History object gets a feedback with required information leaked from a victim site. It is important to mention that the redirect command can appear explicitly in the code, or can be completed by the operational environment.
Attack Vector:
- Create IFRAME with src=Page B
- Remember the current value of history.length
- Change src of IFRAME to Page A
- If the value of history.length is the same, then the CONDITION is TRUE
Login Detection
The following demo for IE and Facebook can show how one can identify if users are currently using facebook: “Am I using Facebook?”
Cross-Site Information Inference
It is possible to inference sensitive information from a page on a different origin, if it implements a conditional redirect. Suppose that in an HR application which is not publically accessible, a legal user can search employees by name, salary and other criteria. If the search has no results, a redirect command is then executed to a “Not Found” page. By submitting the following URL:
and observing the NotFound redirection, attackers can inference sensitive information about a worker’s salary.
This can be done by using the following attack vector:
- Create IFRAME with src=”NotFound.aspx”
- Remember the current value of history.length
- Change src of IFRAME to “SearchEmployee.aspx?name=Jon\&SalaryFrom=3000\&SalaryTo=3500”
- If the value of history.length remains the same, then your search has no results
By repeating the above attack and trying different values of the salary parameters, an attacker can gather very sensitive salary information of any employee. This is a very serious Cross-Site information leakage. If an application has a functionality like a search page with conditional redirect, then this application is vulnerable to XSHM and essentially it is a similar to a direct exposure to Universal XSS – the application itself is XSS-safe, but running it from a different site inside an IFRAME makes it vulnerable.