profit-margin-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of external data files, such as JSON and CSV exports from ecommerce platforms. This creates a surface for indirect prompt injection if the analyzed data contains adversarial content. However, this is a standard risk for any data analysis tool and is managed by the agent's underlying safety protocols. * Ingestion points: External data files mentioned in the evaluation tasks (inputs/sku_data.json) and platform-specific exports. * Boundary markers: The skill instructions do not explicitly mandate the use of delimiters or warnings to separate data from instructions. * Capability inventory: Involves Python and SQL code generation and execution. * Sanitization: Evaluation criteria do not specifically require sanitization of the input data fields.
- [EXTERNAL_DOWNLOADS]: The documentation mentions several reputable third-party ecommerce applications (e.g., BeProfit, Metorik, Glew.io) and official platform extensions. These references are appropriate for the skill's purpose and involve well-known services within the ecommerce ecosystem.
Audit Metadata