autopoisoner - Web cache poisoning vulnerability scanner.

项目说明

Summary

autoPoisoner is your best companion for automating web cache poisoning detection at scale. The tool comes with the following features :

When autoPoisoner launched against a domain/sub-domain, it will firstly attempt to detect interesting behaviors when adding new headers to the request. If the HTTP response with an added header is remarkably different than without, it would be the sign of a potential unkeyed header, relative to the cache, identified.
autoPoisoner unkeyed parameter detection is based on the three following interesting behaviors :

- Reflection of a *CANARY* in the HTTP response

When adding new HTTP headers to the original HTTP request, a value, called a *CANARY,* would be passed as their value (e.g. X-Rewrite-URL: CANARY, X-Forwarded-Host: CANARY). In case this results to “CANARY” appearing/reflecting in the response, this would be the immediate feedback that the injected header is well understood by the web-application.

- Different Status-Code

As pre stated, interesting behavior detection relies on the comparison between two responses, the respecti of a normal HTTP request, and a modified one with added content. To this regard, response’s status-code is a very interesting element to consider. The tool would therefore also relies on status-code , e.g. from HTTP *200*(OK) to *404(Not Found)*, 200 to *302* (Redirect) or 200 to *501* (Not Implemented).

- HTTP response’s length difference

Finally, if no *canary* reflected in the response, and no status-code changes occurred between the original & modified requests, autoPoisoner would consider the HTTP response’s length difference. As web-pages can slightly change over time, the tool would only trigger if the distance between the two HTTP responses’s length is superior to 25% of the original response’s length. This pourcentage has been chosen to limit, as much as possible, false-positives.

In addition to attempt injecting unkeyed headers to the HTTP request, autoPoisoner will also manipulate the *Host* header by adding a port to it (e.g. Host: company.com:8888, expecting web-application different behavior to it. In fact, port could constitute an unkeyed component, to the cache, potentially resulting from requests with Host: company.com being assigned the same cache data as the ones with Host: company.com:8888. This would cause an issue in cases where the web-application beha when a port is added to the host.

Caches often deal with web content differently, depending on their ability to change over time. In fact, due to their nature, static files such as ja, png, jpg, or ico, would rarely be modified, but still often accessed by all when a page, which make them very relevant to cache. For a specified url it would, therefore, get plausible to observe no cache behavior, whereas this would be the case for .
To address this concern, in case no web cache poisoning is detected for a specified URL, autoPoisoner would automatically attempt to crawl static files through the original page’s source code, and conduct new attacks on them.

---

Usage

This displays help for the tool.

Single URL target with *behavior* mode activated:

Multiple URLs target with and fi threads:

---

Example output

autoPoisoner launched against *PortSwigger*’s web cache poisoning lab (with mode activated):

autoPoisoner launched against Swisscom’s multiple sub-domains:

---

Installation

1 - Clone

2 - Install required library

---

License

autoPoisoner is distributed under MIT License.

项目信息

开发语言：Python

Star 数：77

技术要点

该项目属于安全软件方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕Web cache poisoning vulnerability scanner.展开，提供了可运行的实现，对理解相关安全机制、工具原理与攻防技术具有较高的研究价值。

从技术实现角度看，此类项目的核心在于对底层原理的深入理解与工程化落地。开发者在研究过程中可以结合官方文档与社区讨论，梳理项目的初始化流程、核心模块职责与数据流转，逐步掌握其中的技术细节与设计思路。

使用与学习建议

1. 下载源码压缩包，解压到本地环境。

2. 根据项目类型准备对应的运行环境与依赖。

3. 按照 README 中的步骤完成配置与部署。

4. 在隔离的测试环境中运行验证，避免影响生产系统。

免责声明

本资源来源于开源社区，仅供安全研究、技术学习与合法授权范围内的攻防演练使用，请勿用于任何非法或侵害他人权益的用途。使用本项目产生的一切后果由使用者自行承担。