The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external supplier data (financial, operational, and qualitative metrics) to generate risk reports. This creates an attack surface where malicious instructions could be embedded in the data processed by the agent.
Ingestion points: Data is passed into the SupplierRiskAssessment class methods (e.g., assess_financial_risk, assess_operational_risk) and the RiskMonitoringSystem in SKILL.md.
Boundary markers: None present; the code treats input values as standard data points for scoring logic without explicit delimiters or instructions to ignore embedded prompts.
Capability inventory: The skill includes Python logic for data aggregation, weighted scoring, and plotting using matplotlib. It does not contain network or file-writing capabilities in the provided snippets, though it mentions libraries like requests and beautifulsoup4 as recommended tools.
Sanitization: There is no evidence of sanitization or validation of the input strings (such as supplier_name or factors) before they are interpolated into the final risk report output.

supplier-risk-management