corporate-gifting-outreach
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve researching high-value accounts and contacts to identify relevant gift ideas. This data ingestion creates a surface area for indirect prompt injection if the external data sources contain malicious instructions designed to influence the agent.
- Ingestion points: External research data regarding target accounts and contact individuals (referenced in SKILL.md).
- Boundary markers: Absent; there are no instructions to delimit or ignore embedded commands within the researched data.
- Capability inventory: The skill guidelines do not define sensitive capabilities such as file-system modifications or unauthorized network exfiltration.
- Sanitization: Absent; no validation or filtering of the external research content is specified.
- [EXTERNAL_DOWNLOADS]: The skill recommends the addition of an MCP (Model Context Protocol) tool from a vendor-controlled domain:
https://mcp.manypi.com/mcp. This tool is used to fetch external account research details into the agent context.
Audit Metadata