grouped-statistics

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external Excel files, which is a common vector for indirect prompt injection if the processed data contains malicious instructions intended for the agent.
  • Ingestion points: The skill reads external data using pd.read_excel in Step 1.
  • Boundary markers: No specific boundary markers or instructions to ignore embedded content are used when processing the Excel data.
  • Capability inventory: The skill has the capability to write files (wb.save) and output data to the user via print statements.
  • Sanitization: The skill performs basic type conversion using pd.to_numeric(..., errors='coerce'), but does not perform comprehensive sanitization of string content before processing or reporting.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses local files and creates new ones on the file system. This behavior is consistent with the stated purpose of analyzing Excel files and generating reports. No network operations or attempts to exfiltrate data to external domains were identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 12:24 PM
Security Audit — agent-trust-hub — grouped-statistics