gcp-spark-troubleshooting

High-severity arbitrary code execution risk: user-supplied rule and detail expressions reach unrestricted eval with Python builtins available, allowing code to run with the process's privileges if rules are untrusted. No specific malicious payload, credential harvesting, suspicious domain, or direct data exfiltration is present in the fragment. Treat rule files and custom rules as executable code and accept them only from trusted sources; no malicious URL, domain, IP, or targeted credential is identified.