detecting-data-anomalies
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is intended to process and analyze external datasets from sources like CSV files, databases, and APIs. This ingestion of untrusted external content represents a surface for indirect prompt injection, where malicious instructions could be embedded within the data to influence the AI's analysis or report generation.
- Ingestion points: The documentation for
scripts/data_loader.pyindicates that data is loaded from various external sources into the system for analysis. - Boundary markers: The skill instructions do not define specific delimiters or instructions to the agent to ignore or isolate potential instructions found within the processed data.
- Capability inventory: The skill has the capability to select machine learning algorithms, analyze data points, and generate detailed summary reports.
- Sanitization: There are no explicit data validation or sanitization procedures mentioned to prevent the agent from being influenced by instructions embedded in the input data.
Audit Metadata