Possible command injection detected in user event

Rule Details

Cortex AppSec Rule ID	APPSEC_CICD_146
Category	Input Validation
Severity	CRITICAL

Impact

Cortex Cloud has identified an attempt to carry out a command injection attack. Pipelines are vulnerable to command injection attacks in which arbitrary commands and scripts are passed through user controlled inputs such as pull request title, commit description etc., which are referenced by the pipeline configuration file and executed in the pipeline job. When a pipeline job runs, the poisoned controlled inputs can be interpreted as code, which is then executed on the runner, enabling remote code execution attacks to take place in order to access stored secrets or to read or modify the source code.

Recommended Solution - Buildtime

Verify whether there is an actual attempt to inject code. Validate your findings with the responsible user.

For more information refer to https://docs.github.com/en/actions/security-guides/security-hardening-for-github-actions#good-practices-for-mitigating-script-injection-attacks

To prevent a command injection attack, validate all user controlled inputs and ensure that they do not contain malicious commands or scripts.

It is recommended to use an action that processes user input as an argument, instead of an inline script. Another approach would be to set the value of the expression to an intermediate environment variable.

Insufficient Pipeline Based Access Controls (PBAC) may enable attackers to carry out their attack with a wide scope of permissions. Therefore define the scope of permissions that the runner requires according to the principle of least privilege.