Merge pull request #3522 from porcupineyhairs/pythonXpath

Python : Add Xpath injection query

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

@@ -0,0 +1,30 @@
+<!DOCTYPE qhelp SYSTEM "qhelp.dtd">
+<qhelp>
+    <overview>
+        <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
+        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.
+        </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>
+    </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" />
+        <p> This can be fixed by using a parameterized query as shown below.</p>
+        <sample src="xpathGood.py" />
+    </example>
+    <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.ql

@@ -0,0 +1,35 @@
+/**
+ * @name XPath query built from user-controlled sources
+ * @description Building a XPath query from user-controlled sources is vulnerable to insertion of
+ *              malicious Xpath code by the user.
+ * @kind path-problem
+ * @problem.severity error
+ * @precision high
+ * @id py/xpath-injection
+ * @tags security
+ *       external/cwe/cwe-643
+ */
+
+import python
+import semmle.python.security.Paths
+/* Sources */
+import semmle.python.web.HttpRequest
+/* Sinks */
+import experimental.semmle.python.security.injection.Xpath
+
+class XpathInjectionConfiguration extends TaintTracking::Configuration {
+  XpathInjectionConfiguration() { this = "Xpath injection configuration" }
+
+  override predicate isSource(TaintTracking::Source source) {
+    source instanceof HttpRequestTaintSource
+  }
+
+  override predicate isSink(TaintTracking::Sink sink) {
+    sink instanceof XpathInjection::XpathInjectionSink
+  }
+}
+
+from XpathInjectionConfiguration config, TaintedPathSource src, TaintedPathSink sink
+where config.hasFlowPath(src, sink)
+select sink.getSink(), src, sink, "This Xpath query depends on $@.", src.getSource(),
+  "a user-provided value"

python/ql/src/experimental/CWE-643/xpathBad.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='%s']" % value)
+
+
+urlpatterns = [
+    path('a', a)
+]

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

@@ -0,0 +1,115 @@
+/**
+ * Provides class and predicates to track external data that
+ * may represent malicious xpath query objects.
+ *
+ * This module is intended to be imported into a taint-tracking query
+ * to extend `TaintKind` and `TaintSink`.
+ */
+
+import python
+import semmle.python.dataflow.TaintTracking
+import semmle.python.web.HttpRequest
+
+/** Models Xpath Injection related classes and functions */
+module XpathInjection {
+  /** Returns a class value which refers to `lxml.etree` */
+  Value etree() { result = Value::named("lxml.etree") }
+
+  /** Returns a class value which refers to `lxml.etree` */
+  Value libxml2parseFile() { result = Value::named("libxml2.parseFile") }
+
+  /** A generic taint sink that is vulnerable to Xpath injection. */
+  abstract class XpathInjectionSink extends TaintSink { }
+
+  /**
+   * A Sink representing an argument to the `etree.Xpath` call.
+   *
+   *    from lxml import etree
+   *    root = etree.XML("<xmlContent>")
+   *    find_text = etree.XPath("`sink`")
+   */
+  private class EtreeXpathArgument extends XpathInjectionSink {
+    override string toString() { result = "lxml.etree.Xpath" }
+
+    EtreeXpathArgument() {
+      exists(CallNode call | call.getFunction().(AttrNode).getObject("XPath").pointsTo(etree()) |
+        call.getArg(0) = this
+      )
+    }
+
+    override predicate sinks(TaintKind kind) { kind instanceof ExternalStringKind }
+  }
+
+  /**
+   * A Sink representing an argument to the `etree.EtXpath` call.
+   *
+   *    from lxml import etree
+   *    root = etree.XML("<xmlContent>")
+   *    find_text = etree.EtXPath("`sink`")
+   */
+  private class EtreeETXpathArgument extends XpathInjectionSink {
+    override string toString() { result = "lxml.etree.ETXpath" }
+
+    EtreeETXpathArgument() {
+      exists(CallNode call | call.getFunction().(AttrNode).getObject("ETXPath").pointsTo(etree()) |
+        call.getArg(0) = this
+      )
+    }
+
+    override predicate sinks(TaintKind kind) { kind instanceof ExternalStringKind }
+  }
+
+  /**
+   * A Sink representing an argument to the `xpath` call to a parsed xml document.
+   *
+   *    from lxml import etree
+   *    from io import StringIO
+   *    f = StringIO('<foo><bar></bar></foo>')
+   *    tree = etree.parse(f)
+   *    r = tree.xpath('`sink`')
+   */
+  private class ParseXpathArgument extends XpathInjectionSink {
+    override string toString() { result = "lxml.etree.parse.xpath" }
+
+    ParseXpathArgument() {
+      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
+      )
+    }
+
+    override predicate sinks(TaintKind kind) { kind instanceof ExternalStringKind }
+  }
+
+  /**
+   * A Sink representing an argument to the `xpathEval` call to a parsed libxml2 document.
+   *
+   *    import libxml2
+   *    tree = libxml2.parseFile("file.xml")
+   *    r = tree.xpathEval('`sink`')
+   */
+  private class ParseFileXpathEvalArgument extends XpathInjectionSink {
+    override string toString() { result = "libxml2.parseFile.xpathEval" }
+
+    ParseFileXpathEvalArgument() {
+      exists(
+        CallNode parseCall, CallNode xpathCall, ControlFlowNode obj, Variable var, AssignStmt assign
+      |
+        parseCall.getFunction().(AttrNode).pointsTo(libxml2parseFile()) and
+        assign.getValue().(Call).getAFlowNode() = parseCall and
+        xpathCall.getFunction().(AttrNode).getObject("xpathEval") = obj and
+        var.getAUse() = obj and
+        assign.getATarget() = var.getAStore() and
+        xpathCall.getArg(0) = this
+      )
+    }
+
+    override predicate sinks(TaintKind kind) { kind instanceof ExternalStringKind }
+  }
+}

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

@@ -0,0 +1 @@
+semmle-extractor-options: --max-import-depth=3 -p ../../query-tests/Security/lib/

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

@@ -0,0 +1,38 @@
+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:11:18:11:29 | dict of externally controlled string | xpathFlow.py:11:18:11:44 | externally controlled string |
+| xpathFlow.py:11:18:11:29 | dict of externally controlled string | xpathFlow.py:11:18:11:44 | externally controlled string |
+| xpathFlow.py:11:18:11:44 | externally controlled string | xpathFlow.py:14:20:14:29 | externally controlled string |
+| xpathFlow.py:11:18:11:44 | externally controlled string | xpathFlow.py:14:20:14:29 | externally controlled string |
+| xpathFlow.py:20:18:20:29 | dict of externally controlled string | xpathFlow.py:20:18:20:44 | externally controlled string |
+| xpathFlow.py:20:18:20:29 | dict of externally controlled string | xpathFlow.py:20:18:20:44 | externally controlled string |
+| xpathFlow.py:20:18:20:44 | externally controlled string | xpathFlow.py:23:29:23:38 | externally controlled string |
+| xpathFlow.py:20:18:20:44 | externally controlled string | xpathFlow.py:23:29:23:38 | externally controlled string |
+| xpathFlow.py:30:18:30:29 | dict of externally controlled string | xpathFlow.py:30:18:30:44 | externally controlled string |
+| xpathFlow.py:30:18:30:29 | dict of externally controlled string | xpathFlow.py:30:18:30:44 | externally controlled string |
+| xpathFlow.py:30:18:30:44 | externally controlled string | xpathFlow.py:32:29:32:38 | externally controlled string |
+| xpathFlow.py:30:18:30:44 | externally controlled string | xpathFlow.py:32:29:32:38 | externally controlled string |
+| xpathFlow.py:39:18:39:29 | dict of externally controlled string | xpathFlow.py:39:18:39:44 | externally controlled string |
+| xpathFlow.py:39:18:39:29 | dict of externally controlled string | xpathFlow.py:39:18:39:44 | externally controlled string |
+| xpathFlow.py:39:18:39:44 | externally controlled string | xpathFlow.py:41:31:41:40 | externally controlled string |
+| xpathFlow.py:39:18:39:44 | externally controlled string | xpathFlow.py:41:31:41:40 | externally controlled string |
+| xpathFlow.py:47:18:47:29 | dict of externally controlled string | xpathFlow.py:47:18:47:44 | externally controlled string |
+| xpathFlow.py:47:18:47:29 | dict of externally controlled string | xpathFlow.py:47:18:47:44 | externally controlled string |
+| xpathFlow.py:47:18:47:44 | externally controlled string | xpathFlow.py:49:29:49:38 | externally controlled string |
+| xpathFlow.py:47:18:47:44 | externally controlled string | xpathFlow.py:49:29:49:38 | 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:14:20:14:29 | xpathQuery | xpathFlow.py:11:18:11:29 | dict of externally controlled string | xpathFlow.py:14:20:14:29 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:11:18:11:29 | Attribute | a user-provided value |
+| xpathFlow.py:23:29:23:38 | xpathQuery | xpathFlow.py:20:18:20:29 | dict of externally controlled string | xpathFlow.py:23:29:23:38 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:20:18:20:29 | Attribute | a user-provided value |
+| xpathFlow.py:32:29:32:38 | xpathQuery | xpathFlow.py:30:18:30:29 | dict of externally controlled string | xpathFlow.py:32:29:32:38 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:30:18:30:29 | Attribute | a user-provided value |
+| xpathFlow.py:41:31:41:40 | xpathQuery | xpathFlow.py:39:18:39:29 | dict of externally controlled string | xpathFlow.py:41:31:41:40 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:39:18:39:29 | Attribute | a user-provided value |
+| xpathFlow.py:49:29:49:38 | xpathQuery | xpathFlow.py:47:18:47:29 | dict of externally controlled string | xpathFlow.py:49:29:49:38 | externally controlled string | This Xpath query depends on $@. | xpathFlow.py:47:18:47:29 | Attribute | a user-provided value |

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

@@ -0,0 +1,40 @@
+from lxml import etree
+from io import StringIO
+
+
+def a():
+    f = StringIO('<foo><bar></bar></foo>')
+    tree = etree.parse(f)
+    r = tree.xpath('/foo/bar')
+
+
+def b():
+    root = etree.XML("<root><a>TEXT</a></root>")
+    find_text = etree.XPath("//text()")
+    text = find_text(root)[0]
+
+
+def c():
+    root = etree.XML("<root><a>TEXT</a></root>")
+    find_text = etree.XPath("//text()", smart_strings=False)
+    text = find_text(root)[0]
+
+
+def d():
+    root = etree.XML("<root><a>TEXT</a></root>")
+    find_text = find = etree.ETXPath("//{ns}b")
+    text = find_text(root)[0]
+
+
+def e():
+    import libxml2
+    doc = libxml2.parseFile('xpath_injection/credential.xml')
+    results = doc.xpathEval('sink')
+
+
+if __name__ == "__main__":
+    a()
+    b()
+    c()
+    d()
+    e()

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

@@ -0,0 +1 @@
+experimental/CWE-643/xpath.ql

python/ql/test/experimental/CWE-643/xpathBad.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='%s']" % value)
+
+
+urlpatterns = [
+    path('a', a)
+]

python/ql/test/experimental/CWE-643/xpathFlow.py

@@ -0,0 +1,49 @@
+from io import StringIO
+from flask import Flask, request
+
+app = Flask(__name__)
+
+
+@app.route("/xpath1")
+def xpath1():
+    from lxml import etree
+
+    xpathQuery = request.args.get('xml', '')
+    f = StringIO('<foo><bar></bar></foo>')
+    tree = etree.parse(f)
+    r = tree.xpath(xpathQuery)
+
+
+@app.route("/xpath2")
+def xpath2():
+    from lxml import etree
+    xpathQuery = request.args.get('xml', '')
+
+    root = etree.XML("<root><a>TEXT</a></root>")
+    find_text = etree.XPath(xpathQuery)
+    text = find_text(root)[0]
+
+
+@app.route("/xpath3")
+def xpath3():
+    from lxml import etree
+    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 xpath4():
+    from lxml import etree
+    xpathQuery = request.args.get('xml', '')
+    root = etree.XML("<root><a>TEXT</a></root>")
+    find_text = etree.ETXPath(xpathQuery)
+    text = find_text(root)[0]
+
+@app.route("/xpath5")
+def xpath5():
+    import libxml2
+    xpathQuery = request.args.get('xml', '')
+    doc = libxml2.parseFile('xpath_injection/credential.xml')
+    results = doc.xpathEval(xpathQuery)

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

@@ -0,0 +1,12 @@
+| xpath.py:8:20:8:29 | lxml.etree.parse.xpath | externally controlled string |
+| xpath.py:13:29:13:38 | lxml.etree.Xpath | externally controlled string |
+| xpath.py:19:29:19:38 | lxml.etree.Xpath | externally controlled string |
+| xpath.py:25:38:25:46 | lxml.etree.ETXpath | externally controlled string |
+| xpath.py:32:29:32:34 | libxml2.parseFile.xpathEval | externally controlled string |
+| xpathBad.py:13:20:13:43 | lxml.etree.parse.xpath | externally controlled string |
+| xpathFlow.py:14:20:14:29 | lxml.etree.parse.xpath | externally controlled string |
+| xpathFlow.py:23:29:23:38 | lxml.etree.Xpath | externally controlled string |
+| xpathFlow.py:32:29:32:38 | lxml.etree.Xpath | externally controlled string |
+| xpathFlow.py:41:31:41:40 | lxml.etree.ETXpath | externally controlled string |
+| xpathFlow.py:49:29:49:38 | libxml2.parseFile.xpathEval | externally controlled string |
+| xpathGood.py:13:20:13:37 | lxml.etree.parse.xpath | externally controlled string |

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

@@ -0,0 +1,6 @@
+import python
+import experimental.semmle.python.security.injection.Xpath
+
+from XpathInjection::XpathInjectionSink sink, TaintKind kind
+where sink.sinks(kind)
+select sink, kind

python/ql/test/query-tests/Security/lib/libxml2/__init__.py

@@ -0,0 +1,10 @@
+def parseFile(filename):
+    return xmlDoc(_obj=None)
+
+
+class xmlDoc(Object):
+    def __init__(self, _obj=None):
+        pass
+
+    def xpathEval(self, expr):
+        pass

python/ql/test/query-tests/Security/lib/lxml/etree/__init__.py

@@ -0,0 +1,37 @@
+class _ElementTree(object):
+    def xpath(self, _path, namespaces=None, extensions=None, smart_strings=True, **_variables):
+        pass
+
+    def xslt(self, _xslt, extensions=None, access_control=None, **_kw):
+        pass
+
+
+class ETXPath(object):
+    def __init__(self, path, extensions=None, regexp=True, smart_strings=True):
+        pass
+
+
+class XPath(object):
+    def __init__(self, path, namespaces=None, extensions=None, regexp=True, smart_strings=True):
+        pass
+
+
+class XSLT(object):
+    def __init__(self, xslt_input, extensions=None, regexp=True, access_control=None):
+        pass
+
+
+def parse(self, parser=None, base_url=None):
+    return _ElementTree()
+
+
+def fromstring(self, text, parser=None, base_url=None):
+    pass
+
+
+def fromstringlist(self, strings, parser=None):
+    pass
+
+
+def XML(self, text, parser=None, base_url=None):
+    pass