using-superpowers

This document is a procedural/workflow policy that enforces unconditional invocation of external 'skills' whenever there is any chance they apply. It does not itself perform network I/O or contain malware-like code, but it substantially increases supply-chain and operational risk by compelling frequent, low-threshold loading and execution of third-party skill code while forbidding safer review mechanisms. Treat this policy as high-risk operational guidance: require vetting, explicit permission, sandboxing, and the ability to review skill content prior to execution to mitigate credential theft, data exfiltration, and other side effects.