unrestraint-packs

SUSPICIOUS. The skill’s core purpose is to help an agent locate and apply jailbreak/NSFW prompt packs that bypass normal safety behavior. It does not show clear credential theft or exfiltration, but its functionality is fundamentally about subverting model controls, and it references persistent prompt-file integration across multiple agents. Same-account GitHub sourcing lowers publisher-mismatch concern, yet the overall skill remains high-risk due to intentional safety bypass behavior.