skills/santosomar/general-secure-coding-agent-skills/python-regression-test-generator/Gen Agent Trust Hub
python-regression-test-generator
Warn
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill identifies a workflow that uses
pickle.load()to ingest test cases from a file (regression_cases.pkl). This is a dynamic execution risk because thepicklemodule is insecure and can execute arbitrary code during deserialization if the input file is tampered with or maliciously constructed.\n- [DATA_EXFILTRATION]: The skill describes capturing real inputs by monkey-patching functions in staging or production environments. Although it includes guidance to sanitize captured inputs (e.g., removing customer emails), the automated capture of live traffic naturally risks the exposure of PII or other sensitive system data if sanitization is insufficient or improperly implemented.
Audit Metadata