onboarding

SUSPICIOUS. The skill is broadly aligned with legitimate LaunchDarkly onboarding and uses mostly official LaunchDarkly endpoints and normal SDK/package-manager installs, so it is not fundamentally malicious. However, its footprint is fairly expansive for onboarding: it installs transitive skills through a third-party CLI, may fetch and write real credentials, edits agent MCP config, starts local services, and emphasizes quiet execution. Those compounding trust and autonomy signals make it higher risk than a simple documentation or setup skill, though not incompatible with its stated purpose.