outcome-learning (family learning-loop member)

Remember what works for this user across sessions, while keeping every byte of that memory in their workspace and out of the cloud. This is the learning-loop member of the job-hunter family: the read side of the per-user memory layer that keeps the family from re-litigating settled preferences and that surfaces durable patterns — but only ones the user has confirmed.

See references/learning-loop-guide.md for the full design.

When to use this skill

Use this skill when the user wants to learn from outcomes or inspect the per-user memory: