Stronger Anti Cross-Site Scripting (XSS) Filter for Java Web Apps

XSSRequestWrapper.java

package pt.impresa.iweb.filters.request;

import java.io.IOException;
import java.text.Normalizer;
import java.util.Collections;
import java.util.Enumeration;
import java.util.HashMap;
import java.util.Map;
import java.util.Set;
import java.util.regex.Pattern;

import javax.servlet.Filter;
import javax.servlet.FilterChain;
import javax.servlet.FilterConfig;
import javax.servlet.ServletException;
import javax.servlet.ServletRequest;
import javax.servlet.ServletResponse;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletRequestWrapper;

/**
 * Filters Http requests and removes malicious characters/strings
 * (i.e. XSS) from the Query String
 */
public class XSSPreventionFilter implements Filter {
	class XSSRequestWrapper extends HttpServletRequestWrapper {

		private Map<String, String[]> sanitizedQueryString;
		
		public XSSRequestWrapper(HttpServletRequest request) {
			super(request);
		}
		
		//QueryString overrides
		
		@Override
		public String getParameter(String name) {
			String parameter = null;
			String[] vals = getParameterMap().get(name); 
			
			if (vals != null && vals.length > 0) {
				parameter = vals[0];
			}
			
			return parameter;
		}

		@Override
		public String[] getParameterValues(String name) {
			return getParameterMap().get(name);
		}
		
		@Override
		public Enumeration<String> getParameterNames() {
			return Collections.enumeration(getParameterMap().keySet());
		}

		@SuppressWarnings("unchecked")
		@Override
		public Map<String,String[]> getParameterMap() {
			if(sanitizedQueryString == null) {
				Map<String, String[]> res = new HashMap<String, String[]>();
				Map<String, String[]> originalQueryString = super.getParameterMap();
				if(originalQueryString!=null) {
					for (String key : (Set<String>) originalQueryString.keySet()) {
						String[] rawVals = originalQueryString.get(key);
						String[] snzVals = new String[rawVals.length];
						for (int i=0; i < rawVals.length; i++) {
							snzVals[i] = stripXSS(rawVals[i]);
							System.out.println("Sanitized: " + rawVals[i] + " to " + snzVals[i]);
						}
						res.put(stripXSS(key), snzVals);
					}
				}
				sanitizedQueryString = res;
			}
			return sanitizedQueryString;
		}

		//TODO: Implement support for headers and cookies (override getHeaders and getCookies)
		
		/**
		 * Removes all the potentially malicious characters from a string
		 * @param value the raw string
		 * @return the sanitized string
		 */
		private String stripXSS(String value) {
			String cleanValue = null;
			if (value != null) {
				cleanValue = Normalizer.normalize(value, Normalizer.Form.NFD);

				// Avoid null characters
				cleanValue = cleanValue.replaceAll("\0", "");
				
				// Avoid anything between script tags
				Pattern scriptPattern = Pattern.compile("<script>(.*?)</script>", Pattern.CASE_INSENSITIVE);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
		 
				// Avoid anything in a src='...' type of expression
				scriptPattern = Pattern.compile("src[\r\n]*=[\r\n]*\\\'(.*?)\\\'", Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");

				scriptPattern = Pattern.compile("src[\r\n]*=[\r\n]*\\\"(.*?)\\\"", Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
				
				// Remove any lonesome </script> tag
				scriptPattern = Pattern.compile("</script>", Pattern.CASE_INSENSITIVE);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");

				// Remove any lonesome <script ...> tag
				scriptPattern = Pattern.compile("<script(.*?)>", Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");

				// Avoid eval(...) expressions
				scriptPattern = Pattern.compile("eval\\((.*?)\\)", Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
				
				// Avoid expression(...) expressions
				scriptPattern = Pattern.compile("expression\\((.*?)\\)", Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
				
				// Avoid javascript:... expressions
				scriptPattern = Pattern.compile("javascript:", Pattern.CASE_INSENSITIVE);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
				
				// Avoid vbscript:... expressions
				scriptPattern = Pattern.compile("vbscript:", Pattern.CASE_INSENSITIVE);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
				
				// Avoid onload= expressions
				scriptPattern = Pattern.compile("onload(.*?)=", Pattern.CASE_INSENSITIVE | Pattern.MULTILINE | Pattern.DOTALL);
				cleanValue = scriptPattern.matcher(cleanValue).replaceAll("");
			}
			return cleanValue;
		}
	}

	@Override
	public void destroy() {
		System.out.println("XSSPreventionFilter: destroy()");
	}

	@Override
	public void doFilter(ServletRequest request, ServletResponse response, 
			FilterChain chain) throws IOException, ServletException {
		XSSRequestWrapper wrapper = new XSSRequestWrapper((HttpServletRequest)request);
		chain.doFilter(wrapper, response);
	}

	@Override
	public void init(FilterConfig filterConfig) throws ServletException {
		System.out.println("XSSPreventionFilter: init()");
	}
}

For handling Unicode character like : \u003cscript\u003e\u003c/script\u003e ; you can add below 3 steps using apache commons lang library.
value=StringEscapeUtils.unescapeJava(value);
value = new URLCodec().decode(value);

In stripXSS method you are compiling several regexps each time you filter HTTP request. Pattern.compile is quite expensive operation from performance perspective, so I strongly suggest to compile all Patterns once at the start of application by making them private static final variables.

Please also see this Q&A here: https://softwareengineering.stackexchange.com/questions/216320/java-regex-patterns-compile-time-constants-or-instance-members

@sombra-yuriy you're right about the Pattern.compile(). It should be done outside the stripXSS method.

This code snippet is completely outdated and should not be used. I'm keeping it here for historical reasons xD. However,

@madoke will be have new updated version of this code snippet

@jibz08 not really. There is probably a better way to protect your website.