markdown-optimization
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides legitimate retail strategy guidance and mathematical models for price optimization. The Python code blocks are self-contained examples that process provided data using standard libraries (Pandas, NumPy, SciPy) without making network requests, accessing sensitive files, or executing arbitrary system commands.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process retail data (historical pricing, sales, inventory) to generate optimization reports. While this creates a data ingestion surface, the risk is negligible as the processing logic is strictly mathematical and does not involve interpreting natural language instructions embedded within the data.
- Ingestion points: SKILL.md (via Python scripts
PriceElasticityAnalyzer,MarkdownScheduler,MarkdownOptimizer, andStoreMarkdownManagerwhich process input DataFrames). - Boundary markers: Not explicitly defined for the data input fields.
- Capability inventory: Performs mathematical optimization and logic using
scipy.optimize,numpy, andpandas. No file writing, network operations, or dynamic code evaluation (eval/exec) are present. - Sanitization: Standard DataFrame operations are used which inherently limit the interpretation of data as instructions.
Audit Metadata