executing-plans

SUSPICIOUS: the skill’s core purpose is coherent for plan execution, but it relies on transitive skill installation/invocation and permits autonomous push/PR actions. There is no direct evidence of credential theft, exfiltration, remote payload execution, or malicious concealment in this skill text; the main risk is inherited trust in unverified `razorback:*` and `Julie` components.