In this paper, we propose a client-oriented web alteration detection system that uses the changed state of links between the past and current versions of a browsing web page. Some measures against malicious web page alterations have been developed, such as URL-blacklist based access control. The limitations of such measures, however, are the coverage and timeliness. It is very difficult to capture all maliciously altered web page data on the Internet. A time delay from when an attacker alters a web page to when a browser blocks access to the web page is unavoidable. We focus on a page alteration such as a code injection or a page modification without a change in the visual layout. Such an alteration may lead users to experience further security threats. To detect altered web pages, whenever a user views a web page, our system extracts the link-based feature data from the page and stores it in a database. In addition, if the database has the feature data of a browsing web page from a previous access time, the system extracts the change in state of all links on the web page based on both the previous and current page content. Moreover, the results are provided to the users through a visual representation. Our system assists web-browsing users to remain aware of malicious alterations to a browsing web page. We believe that our system can engage web-browsing users to monitor web page alterations.