include sugestions from review

Author: Porcupiney Hairs

python/ql/src/experimental/CWE-643/xpath.qhelp

@@ -1,32 +1,30 @@
 <!DOCTYPE qhelp SYSTEM "qhelp.dtd">
 <qhelp>
     <overview>
-        Using user-supplied information to construct an XPath query for XML data can 
+        <p>
+        Using user-supplied information to construct an XPath query for XML data can
         result in an XPath injection flaw. By sending intentionally malformed information,
-        an attacker can access data that he may not normally have access to. 
-        He/She may even be able to elevate his privileges on the web site if the XML data 
+        an attacker can access data that he may not normally have access to.
+        He/She may even be able to elevate his privileges on the web site if the XML data
         is being used for authentication (such as an XML based user file).
+        </p>
     </overview>
     <recommendation>
         <p>
-        XPath injection can be prevented using parameterized XPath interface or escaping the user input to make it safe to include in a dynamically constructed query. 
-        If you are using quotes to terminate untrusted input in a dynamically constructed XPath query, then you need to escape that quote in the untrusted input to ensure the untrusted data can’t try to break out of that quoted context. 
+        XPath injection can be prevented using parameterized XPath interface or escaping the user input to make it safe to include in a dynamically constructed query.
+        If you are using quotes to terminate untrusted input in a dynamically constructed XPath query, then you need to escape that quote in the untrusted input to ensure the untrusted data can’t try to break out of that quoted context.
         </p>
         <p>
         Another better mitigation option is to use a precompiled XPath query. Precompiled XPath queries are already preset before the program executes, rather than created on the fly after the user’s input has been added to the string. This is a better route because you don’t have to worry about missing a character that should have been escaped.
         </p>
-        <example>
-
-            <p>In the example below, the xpath query is controlled by the user and hence leads to a vulnerability.</p>
-            
-            <sample src="xpath.py" />
-        </example>
-        <references>
-            <li>OWASP XPath injection : <a href="https://owasp.org/www-community/attacks/XPATH_Injection"></a>/>> </li>
-        </references>
-            
-            
     </recommendation>
-    
-
+    <example>
+        <p>In the example below, the xpath query is controlled by the user and hence leads to a vulnerability.</p>
+        <sample src="xpathBad.py" />
+    </example>
+        <p> This can be fixed by using a parameterized query as shown below.</p>
+        <sample src="xpathGood.py" />
+    <references>
+        <li>OWASP XPath injection : <a href="https://owasp.org/www-community/attacks/XPATH_Injection"></a>/>> </li>
+    </references>
 </qhelp>

python/ql/src/experimental/CWE-643/xpath.py renamed to python/ql/src/experimental/CWE-643/xpathBad.py

@@ -7,10 +7,10 @@
 
 
 def a(request):
-    xpathQuery = request.GET['xpath']
+    value = request.GET['xpath']
     f = StringIO('<foo><bar></bar></foo>')
     tree = etree.parse(f)
-    r = tree.xpath(xpathQuery)
+    r = tree.xpath("/tag[@id='%s']" % value)
 
 
 urlpatterns = [

python/ql/src/experimental/CWE-643/xpathGood.py

@@ -0,0 +1,18 @@
+from lxml import etree
+from io import StringIO
+
+from django.urls import path
+from django.http import HttpResponse
+from django.template import Template, Context, Engine, engines
+
+
+def a(request):
+    value = request.GET['xpath']
+    f = StringIO('<foo><bar></bar></foo>')
+    tree = etree.parse(f)
+    r = tree.xpath("/tag[@id=$tagid]", tagid=value)
+
+
+urlpatterns = [
+    path('a', a)
+]

python/ql/src/experimental/semmle/python/security/injection/Xpath.qll

@@ -7,7 +7,7 @@
  */
 
 import python
-import semmle.python.security.TaintTracking
+import semmle.python.dataflow.TaintTracking
 import semmle.python.web.HttpRequest
 
 /** Models Xpath Injection related classes and functions */
@@ -29,11 +29,7 @@ module XpathInjection {
     override string toString() { result = "lxml.etree.Xpath" }
 
     EtreeXpathArgument() {
-      exists(CallNode call, AttrNode atr |
-        atr = etree().getAReference().getASuccessor() and
-        atr.getName() = "XPath" and
-        atr = call.getFunction()
-      |
+      exists(CallNode call | call.getFunction().(AttrNode).getObject("XPath").pointsTo(etree()) |
         call.getArg(0) = this
       )
     }
@@ -52,11 +48,7 @@ module XpathInjection {
     override string toString() { result = "lxml.etree.ETXpath" }
 
     EtreeETXpathArgument() {
-      exists(CallNode call, AttrNode atr |
-        atr = etree().getAReference().getASuccessor() and
-        atr.getName() = "ETXPath" and
-        atr = call.getFunction()
-      |
+      exists(CallNode call | call.getFunction().(AttrNode).getObject("ETXPath").pointsTo(etree()) |
         call.getArg(0) = this
       )
     }
@@ -77,17 +69,15 @@ module XpathInjection {
     override string toString() { result = "lxml.etree.parse.xpath" }
 
     ParseXpathArgument() {
-      exists(CallNode parseCall, AttrNode parse, string s |
-        parse = etree().getAReference().getASuccessor() and
-        parse.getName() = "parse" and
-        parse = parseCall.getFunction() and
-        exists(CallNode xpathCall, AttrNode xpath |
-          xpath = parseCall.getASuccessor*() and
-          xpath.getName() = "xpath" and
-          xpath = xpathCall.getFunction() and
-          s = xpath.getName() and
-          this = xpathCall.getArg(0)
-        )
+      exists(
+        CallNode parseCall, CallNode xpathCall, ControlFlowNode obj, Variable var, AssignStmt assign
+      |
+        parseCall.getFunction().(AttrNode).getObject("parse").pointsTo(etree()) and
+        assign.getValue().(Call).getAFlowNode() = parseCall and
+        xpathCall.getFunction().(AttrNode).getObject("xpath") = obj and
+        var.getAUse() = obj and
+        assign.getATarget() = var.getAStore() and
+        xpathCall.getArg(0) = this
       )
     }
 

python/ql/test/experimental/CWE-643/XpathLibTests/options

@@ -1,4 +1,14 @@
 edges
+| xpathBad.py:9:7:9:13 | django.request.HttpRequest | xpathBad.py:10:13:10:19 | django.request.HttpRequest |
+| xpathBad.py:9:7:9:13 | django.request.HttpRequest | xpathBad.py:10:13:10:19 | django.request.HttpRequest |
+| xpathBad.py:10:13:10:19 | django.request.HttpRequest | xpathBad.py:10:13:10:23 | django.http.request.QueryDict |
+| xpathBad.py:10:13:10:19 | django.request.HttpRequest | xpathBad.py:10:13:10:23 | django.http.request.QueryDict |
+| xpathBad.py:10:13:10:23 | django.http.request.QueryDict | xpathBad.py:10:13:10:32 | externally controlled string |
+| xpathBad.py:10:13:10:23 | django.http.request.QueryDict | xpathBad.py:10:13:10:32 | externally controlled string |
+| xpathBad.py:10:13:10:32 | externally controlled string | xpathBad.py:13:39:13:43 | externally controlled string |
+| xpathBad.py:10:13:10:32 | externally controlled string | xpathBad.py:13:39:13:43 | externally controlled string |
+| xpathBad.py:13:39:13:43 | externally controlled string | xpathBad.py:13:20:13:43 | externally controlled string |
+| xpathBad.py:13:39:13:43 | externally controlled string | xpathBad.py:13:20:13:43 | externally controlled string |
 | xpathFlow.py:10:18:10:29 | dict of externally controlled string | xpathFlow.py:10:18:10:44 | externally controlled string |
 | xpathFlow.py:10:18:10:29 | dict of externally controlled string | xpathFlow.py:10:18:10:44 | externally controlled string |
 | xpathFlow.py:10:18:10:44 | externally controlled string | xpathFlow.py:13:20:13:29 | externally controlled string |
@@ -13,10 +23,11 @@ edges
 | xpathFlow.py:27:18:27:44 | externally controlled string | xpathFlow.py:29:29:29:38 | externally controlled string |
 | xpathFlow.py:35:18:35:29 | dict of externally controlled string | xpathFlow.py:35:18:35:44 | externally controlled string |
 | xpathFlow.py:35:18:35:29 | dict of externally controlled string | xpathFlow.py:35:18:35:44 | externally controlled string |
-| xpathFlow.py:35:18:35:44 | externally controlled string | xpathFlow.py:37:38:37:47 | externally controlled string |
-| xpathFlow.py:35:18:35:44 | externally controlled string | xpathFlow.py:37:38:37:47 | externally controlled string |
+| xpathFlow.py:35:18:35:44 | externally controlled string | xpathFlow.py:37:31:37:40 | externally controlled string |
+| xpathFlow.py:35:18:35:44 | externally controlled string | xpathFlow.py:37:31:37:40 | externally controlled string |
 #select
+| xpathBad.py:13:20:13:43 | BinaryExpr | xpathBad.py:9:7:9:13 | django.request.HttpRequest | xpathBad.py:13:20:13:43 | externally controlled string | This Xpath query depends on $@. | xpathBad.py:9:7:9:13 | request | a user-provided value |
 | xpathFlow.py:13:20:13:29 | xpathQuery | xpathFlow.py:10:18:10:29 | dict of externally controlled string | xpathFlow.py:13:20:13:29 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:10:18:10:29 | Attribute | a user-provided value |
 | xpathFlow.py:21:29:21:38 | xpathQuery | xpathFlow.py:18:18:18:29 | dict of externally controlled string | xpathFlow.py:21:29:21:38 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:18:18:18:29 | Attribute | a user-provided value |
 | xpathFlow.py:29:29:29:38 | xpathQuery | xpathFlow.py:27:18:27:29 | dict of externally controlled string | xpathFlow.py:29:29:29:38 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:27:18:27:29 | Attribute | a user-provided value |
-| xpathFlow.py:37:38:37:47 | xpathQuery | xpathFlow.py:35:18:35:29 | dict of externally controlled string | xpathFlow.py:37:38:37:47 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:35:18:35:29 | Attribute | a user-provided value |
+| xpathFlow.py:37:31:37:40 | xpathQuery | xpathFlow.py:35:18:35:29 | dict of externally controlled string | xpathFlow.py:37:31:37:40 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:35:18:35:29 | Attribute | a user-provided value |

python/ql/test/experimental/CWE-643/XpathLibTests/xpathSinks.expected

@@ -0,0 +1,18 @@
+from lxml import etree
+from io import StringIO
+
+from django.urls import path
+from django.http import HttpResponse
+from django.template import Template, Context, Engine, engines
+
+
+def a(request):
+    value = request.GET['xpath']
+    f = StringIO('<foo><bar></bar></foo>')
+    tree = etree.parse(f)
+    r = tree.xpath("/tag[@id='%s']" % value)
+
+
+urlpatterns = [
+    path('a', a)
+]

python/ql/test/experimental/CWE-643/xpath.expected

@@ -6,15 +6,15 @@
 
 
 @app.route("/xpath1")
-def a():
+def xpath1():
     xpathQuery = request.args.get('xml', '')
     f = StringIO('<foo><bar></bar></foo>')
     tree = etree.parse(f)
     r = tree.xpath(xpathQuery)
 
 
 @app.route("/xpath2")
-def b():
+def xpath2():
     xpathQuery = request.args.get('xml', '')
 
     root = etree.XML("<root><a>TEXT</a></root>")
@@ -23,16 +23,16 @@ def b():
 
 
 @app.route("/xpath3")
-def c():
+def xpath3():
     xpathQuery = request.args.get('xml', '')
     root = etree.XML("<root><a>TEXT</a></root>")
     find_text = etree.XPath(xpathQuery, smart_strings=False)
     text = find_text(root)[0]
 
 
 @app.route("/xpath4")
-def d():
+def xpath4():
     xpathQuery = request.args.get('xml', '')
     root = etree.XML("<root><a>TEXT</a></root>")
-    find_text = find = etree.ETXPath(xpathQuery)
-    text = find_text(root)[0]
+    find_text = etree.ETXPath(xpathQuery)
+    text = find_text(root)[0]