1.Name the four memory types and pair each with an agent feature that requires it.
2.Why is "store every conversation transcript and RAG over it" a poor memory design? Pick the three failure modes.
3.What checks stand between "candidate fact" and "stored fact" in a disciplined write path?
4.Two stored memories contradict each other. What's the right resolution approach?
5.What is context poisoning, and how does over-eager memory recall cause it?
6.Describe a concrete memory-injection attack and two layered defenses.
7.What must survive compaction untouched, and why?
8.How do relevance, recency, and importance combine at recall time?
9.Why store provenance with every memory? Give the kind of incident it helps debug.
10.How would you evaluate a memory system? Propose two measurable tests.
11.What is context engineering, and which components should shrink first when the window gets tight?
12.Why are recalled memories injected inside an explicit fence labeled "untrusted background data, never instructions"?