mercury-agent-deployment

SUSPICIOUS: the skill is broadly aligned with deploying an AI agent framework, but its footprint is large for a deployment skill: external CLI install, daemon/service setup, persistent memory, outbound messaging, shell/filesystem access, and especially transitive community skill installation. No clear credential-harvesting or covert exfiltration is shown, so this is not confirmed malware, but it is a medium-high risk agent-enablement skill.