famou-data-analysis

Pass

Audited by Gen Agent Trust Hub on Apr 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a legitimate set of instructions for data analysis, focusing on understanding business context, evaluating data quality, and maintaining user alignment.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill references specific file paths (/mnt/user-data/uploads/ and /mnt/user-data/outputs/) which are standard operational directories for processing user-supplied data. No access to sensitive system files or credentials was detected.\n- [INDIRECT_PROMPT_INJECTION]:\n
  • Ingestion points: Reads user-uploaded files (CSV, Excel, datasets) from /mnt/user-data/uploads/.\n
  • Boundary markers: The instructions do not specify explicit delimiters or warnings to ignore instructions embedded within the data files.\n
  • Capability inventory: Performs data reading, transformation, and statistical reporting using standard data science libraries.\n
  • Sanitization: No specific sanitization or filtering of data content is mentioned, though the skill emphasizes verifying actions with the user before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 13, 2026, 05:39 AM