atai-label-free-anomaly-discovery-agent-data-prep

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes sensor data from user-provided CSV files. This ingestion of external data is a potential attack surface for indirect prompt injection, although the skill's operations are limited to local numerical processing.
  • Ingestion points: The scripts read CSV data via pandas.read_csv in lf_prep.py and preflight_raw.py.
  • Boundary markers: No specific delimiters or safety instructions are defined for the input data content.
  • Capability inventory: The skill is restricted to local data manipulation, file reads, and file writes. It does not perform network operations or execute arbitrary shell commands.
  • Sanitization: The skill performs type checking and grid alignment but does not include content-based sanitization for text fields.
  • [EXTERNAL_DOWNLOADS]: The skill requires standard Python packages (numpy, pandas) from official registries, which are well-known and trusted sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 07:26 PM