Happyhacking-K

**Name of the Vulnerable Software and Affected Versions** MaxKB versions prior to 2.0.0 **Description** MaxKB is an open-source AI assistant for enterprise. Prior to version 2.0.0, the sandbox design rules can be bypassed because the software only restricts the execution permissions of files in a specific directory. An attacker can use the `shutil.copy2` method in Python to copy a command to the executable directory, bypassing directory restrictions and enabling reverse shell access. **Recommendations** Update to version 2.0.0 or later.

**Name of the Vulnerable Software and Affected Versions** MaxKB versions prior to 1.10.8-lts **Description** The issue allows attackers to exploit certain files with execution permissions in non-blacklisted directories to carry out attacks, as the Sandbox only restricts the execution permissions of binary files in common directories, such as `/bin`, `/usr/bin`, etc. **Recommendations** For versions prior to 1.10.8-lts, update to version 1.10.8-lts to resolve the issue. As a temporary workaround, consider restricting execution permissions for binary files in non-blacklisted directories to minimize the risk of exploitation.

**Name of the Vulnerable Software and Affected Versions** Apache Airflow Spark Provider versions prior to 4.1.3 **Description** The issue is related to deserialization of untrusted data and inclusion of functionality from an untrusted control sphere. When the Apache Spark provider is installed on an Airflow deployment, an authorized Airflow user can run arbitrary code on the Airflow node by pointing it at a malicious Spark server. Administrators may have provided authorizations to configure Spark hooks without considering this risk, as it was not explicitly mentioned in the documentation prior to version 4.1.3. **Recommendations** To resolve the issue, administrators should review their configurations to ensure that authorization to configure Spark hooks is only provided to fully trusted users. For versions prior to 4.1.3, consider restricting access to the Spark hooks configuration to minimize the risk of exploitation until a patch is available.

**Name of the Vulnerable Software and Affected Versions** Apache Airflow Sqoop Provider versions prior to 4.0.0 **Description** The issue is related to insufficient input validation, which can be exploited by a remote attacker to execute arbitrary code. This can be achieved by passing parameters with connections, making it possible to implement RCE attacks via `sqoop import --connect`, and obtain Airflow server permissions. The attacker needs to be logged in and have authorization to create or edit connections. **Recommendations** To resolve the issue, upgrade to a version that is not affected, specifically version 4.0.0 or later. As a temporary workaround, consider restricting access to the `sqoop import --connect` command to minimize the risk of exploitation. Additionally, limit the ability to create or edit connections to authorized personnel only.

**Name of the Vulnerable Software and Affected Versions** undici version 5.8.0 **Description** The issue is related to CRLF Injection on headers when using unsanitized input as request headers, specifically inside the `content-type` header. This can lead to performing multiple requests in a single API call. For example, using the `content-type` header with unsanitized input like `'application/jsonr r GET /foo2 HTTP/1.1'` can result in two requests: one to `http://localhost:3000/` and another to `http://localhost:3000/foo2`. **Recommendations** For undici version 5.8.0, update to version 5.8.1 to resolve the issue. As a temporary workaround, sanitize input when sending `content-type` headers using user input.