atris@3.69.0
you say what you want in plain words. atris builds it, checks it, and shows you proof.
AI security review
scanned · by lpm-firewall-aiThe package has user-invoked commands that write project agent instructions and link skills into global agent directories. Brain activation requires --yes, and the manifest lists only a prepare hook, so the inspected source does not support blocking the package for install-time agent hijacking.