PRP: Exposed K8s dashboard

[JamesFoxxx]

#457
google/security-testbeds#146

[giacomo-doyensec]

Hi @JamesFoxxx, thanks for your contribution!
I’m running into some issues with the detector. If the callback server returns a 404 or is unavailable, the workflow fails and skips the cleanup phase. Similarly, if the job already exists, the detector fails and again skips cleanup.

Could you update the implementation to handle these edge cases properly? Ideally, the job should use a random name (to avoid clashes with existing ones), and a cleanup action should be added to trigger_code_execution . Thanks!

[JamesFoxxx]

About the random string, I think we need to borrow a random string from the target server itself(like the csrftoken) and append it to the job name, or we can use a constant job name and ignore the matching of job cleanup action response status code and run this action after the fingerprint action( it will be rerun as a cleanup action again). The reason for ignoring the response status code is that if the job with a constant name doesn't exist, it returns 404; if the job is created and exists, then it returns 200.

[JamesFoxxx]

@giacomo-doyensec I implemented an additional job to clean up after the fingerprint action for demonstration.

[giacomo-doyensec]

LGTM - Approved
@tooryx we can merge this and google/security-testbeds#146

Reviewer: Giacomo, Doyensec
Plugin: Exposed K8s Dashboard