linkedin-contrarian-takes

SUSPICIOUS. The core writing behavior is benign and proportionate, but the skill includes transitive installation of another skill and a handoff to a hosted Liftli MCP service. The strongest concern is install trust: the cited `liftli-ai/linkedin-agent-skills` source does not match current official repo evidence (`liftli-ai/skills`), so provenance is inconsistent even though the domains and broader org relationship appear legitimate.