using-superpowers

Overall, the skill/policy is conceptually benign as a governance mechanism but introduces elevated risk surfaces around prompt injection, uncontrolled skill loading, and potential data exposure via the Skill tool. The mandatory invocation flow increases attacker surface area if skills can be loaded from untrusted sources or if the tool interface is abused to pull in arbitrary content. Recommend restricting Skill tool sources to verified, auditable skill registries, adding per-skill consent prompts for external calls, and logging/monitoring skill invocations to detect anomalous behavior.