newly-registered-company-leads
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a shell command to add the ManyPI MCP server (
claude mcp add), which enables external tool integration for lead processing. - [EXTERNAL_DOWNLOADS]: The configuration command references the author's remote domain
mcp.manypi.com. This is a vendor-owned resource used for the skill's intended functionality. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from company registries and websites, which is a surface for indirect injection attacks.
- Ingestion points: Public company registers, VAT lists, and corporate websites mentioned in SKILL.md.
- Boundary markers: No delimiters or protective instructions are specified for handling the external content.
- Capability inventory: The skill utilizes the
manypiMCP tool to interact with the retrieved data. - Sanitization: There is no specification for sanitizing or filtering external input prior to agent processing.
Audit Metadata