stockout-alerts
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to fetch and research external websites and public sources to build a product context, and to ingest inventory and ad spend data via CSV or direct paste.
- Ingestion points: Researching company sites and public sources (SKILL.md); processing CSV or pasted inventory and ad spend data (SKILL.md).
- Boundary markers: The instructions lack explicit boundary markers or directives for the agent to ignore potentially malicious instructions embedded in the external content or CSV data.
- Capability inventory: The skill writes to local files (.agents/product-context.md, ledger), fetches external URLs for research, and interacts with tools via the Model Context Protocol (MCP).
- Sanitization: No specific sanitization or validation logic is defined for the external data before it is processed or written to context files.
- [DATA_EXFILTRATION]: The skill performs network operations to fetch content from external, non-whitelisted domains (such as company websites and public research sources) as part of its 'Before you write' research phase. While intended for legitimate business research, these network calls represent a potential vector for data exposure if sensitive context is included in the requests.
Audit Metadata