list-building

SUSPICIOUS. The skill is largely aligned with its stated email/phone validation purpose, but it pushes a same-org yet unpinned `npx @latest` MCP install and routes user data/API access through ColdIQ's brokered multi-provider service. This is not fundamentally incompatible with the skill's purpose, but the install trust and credential/data forwarding model are stronger than a simple reference skill and warrant medium risk.

SUSPICIOUS-leaning but not malicious: the skill's purpose, capabilities, and data flows are mostly coherent for ABM account selection, and the API endpoint appears official. The main risk is install trust and credential forwarding through an unpinned npx-launched MCP package whose exact first-party install provenance was not clearly verified in the provided evidence.

SUSPICIOUS: The skill’s core dedup purpose is mostly coherent, but it pushes a vendor-specific MCP install and routes credentials through an unpinned npx package whose exact public provenance was not fully verified. Data flows are broadly aligned with list-building, yet the install-trust and credential-forwarding footprint is larger than a simple dedup guidance skill needs.

SUSPICIOUS: The skill’s business purpose is coherent and its ColdIQ data flow is broadly aligned, but it asks users to execute an unpinned npm MCP package and pass an API key into it. With same-brand official domains but unverified package provenance, this is a supply-chain and credential-forwarding risk rather than confirmed malware.