skill-maker

SUSPICIOUS. The skill is coherent with its stated purpose, and the referenced skills CLI appears to be an official same-ecosystem tool distributed via npm/GitHub rather than a random payload. The main issue is transitive trust: it instructs the agent to install other skills into real agent directories, which expands trust to additional skill content and can trigger telemetry. Because the installation workflow is central to the skill's purpose and uses documented official channels, this is better classified as medium-risk transitive trust rather than malicious behavior.