POC详情: 4235d4975e9b300a67f55f880df425e67ec977b7

来源
关联漏洞
标题: Apache Parquet 代码问题漏洞 (CVE-2025-30065)
描述:Apache Parquet是美国阿帕奇(Apache)基金会的一种列式存储格式。可用于 Hadoop 生态系统中的任何项目。 Apache Parquet 1.15.0及之前版本存在代码问题漏洞,该漏洞源于parquet-avro模块的模式解析可能导致执行任意代码。
介绍
# CVE-2025-30065

This repository illustrates how to exploit CVE-2025-30065 and achieve remote class instantiation and  trigger a 
network request from within the victim application i.e the JVM thereby achieving SSRF.

The generated parquet in [Malicious.java](src%2Fmain%2Fjava%2Fcom%2Fevil%2FMalicious.java) assumes that the class [RCEPayload.java](src%2Fmain%2Fjava%2Fcom%2Fevil%2FRCEPayload.java)
is present in the classpath which is not realistic. You can trigger that PoC by executing [Reader.java](src%2Fmain%2Fjava%2Fcom%2Fvictim%2FReader.java).

![image.png](images%2Fimage.png)


[MaliciousSSRF.java](src%2Fmain%2Fjava%2Fcom%2Fevil%2FMaliciousSSRF.java) is more feasible (from an attacker perspective) and triggers a network connection which could be internal or external.
One could also find other gadgets to achieve RCE (what is tricky is to find an instructor accepting a string as arg and leading to a RCE it is not like plain java deser). 
Similarly, PoC can be executed using [ReaderSSRF.java](src%2Fmain%2Fjava%2Fcom%2Fvictim%2FReaderSSRF.java).
![image-ssrf.png](images%2Fimage-ssrf.png)

For more details about the internal of the vulnerability and the fix, you can have a look at my blogpost:  www.deep-kondah.com/parquet-under-fire-a-technical-analysis-of-cve-2025-30065

文件快照

[4.0K] /data/pocs/4235d4975e9b300a67f55f880df425e67ec977b7 ├── [ 147] DISCLAIMER.md ├── [ 83] exploit.html ├── [4.0K] images │   ├── [ 26K] image.png │   └── [ 52K] image-ssrf.png ├── [1.5K] pom.xml ├── [1.3K] README.md └── [4.0K] src └── [4.0K] main └── [4.0K] java └── [4.0K] com ├── [4.0K] evil │   ├── [1.4K] GenerateMaliciousParquet.java │   ├── [1.5K] GenerateMaliciousParquetSSRF.java │   └── [ 509] RCEPayload.java └── [4.0K] victim ├── [ 558] Reader.java └── [ 567] ReaderSSRF.java 7 directories, 11 files
神龙机器人已为您缓存
备注
    1. 建议优先通过来源进行访问。
    2. 如果因为来源失效或无法访问,请发送邮箱到 f.jinxu#gmail.com 索取本地快照(把 # 换成 @)。
    3. 神龙已为您对POC代码进行快照,为了长期维护,请考虑为本地POC付费,感谢您的支持。