dataforseo-mcp-server

The code has a prototype-pollution risk when field paths or traversed input keys are attacker-controlled. The risk arises from dynamic writes to ordinary objects and traversal of inherited properties. No evidence of malware or data exfiltration is present in this fragment.

The workflow has a potentially serious shell-injection issue because the pushed tag is interpolated directly into a shell script. An attacker able to create a matching tag that triggers this workflow may be able to execute commands in the job context. Use an environment variable for github.ref_name and quote its use, or validate the tag before processing. No explicit malicious payload is present in the workflow.

No clear malware is present. The principal security concern is conditional credential disclosure: if an untrusted caller can supply an absolute URL through options.path, the function sends the authorization value to that host. Restrict destinations to the configured API origin or omit credentials for other origins.