Spark 2.5 Arbitrary File Read

2016.11.05
Credit: aj
Risk: Medium
Local: No
Remote: Yes
CVE: N/A
CWE: CWE-200

Hey folks, Spark (sparkjava.com) is a mildly hyped Java micro web framework that also provides functionality to serve static files. Unfortunately, there's no protection against directory traversal attacks and I haven't been able to contact anyone related to the project (after trying 4 people over 2 weeks). As this bug is not that awesome, and fairly trivial to find, please help yourself to some semi-shitty 0-day. If configured, Spark provides two ways to serve static files, either from the classpath via Spark.staticFileLocation() or the filesystem via Spark.externalStaticFileLocation(). ----------------------------------- Classpath Vuln ----------------------------------- Exploit the classpath based vulnerability with something like: curl "http://<host>/..\..\spark\Spark.class" The number of ..\ you need in the path depends on where in the classpath the static file location is configured to be. If you don't have the right amount then you don't get anything, but this is trivially brute forced. The classpath often has juicy things like configuration files and keystores, or you could look for non-framework class files and decompile them. ----------------------------------- Filesystem Vuln ----------------------------------- Try something like: curl "http://<host>/..\..\..\..\..\..\..\..\..\..\..\..\..\..\..\..\..\etc\passwd" I'm sure you can figure out what to do from here. This is tested on Spark 2.5 and git master, it should work on earlier versions. If you need a workaround, don't use Spark to serve static files and move them to another web server. Cheers, aj


Vote for this issue:
50%
50%


 

Thanks for you vote!


 

Thanks for you comment!
Your message is in quarantine 48 hours.

Comment it here.


(*) - required fields.  
{{ x.nick }} | Date: {{ x.ux * 1000 | date:'yyyy-MM-dd' }} {{ x.ux * 1000 | date:'HH:mm' }} CET+1
{{ x.comment }}

Copyright 2024, cxsecurity.com

 

Back to Top