Long-running multi-agent cases accumulate huge transcripts until agents hit context limits and drop critical entities. You need a strategy for context accumulation, retrieval, injection, and compaction across agents. What should you implement?
Select an answer to reveal the explanation.
Short Explanation
Correct answer: B. Context management is luggage limits for agents: pack the structured case state and the snippets that matter, summarize the rest, and don’t lose the passenger names (entities). Dumping the entire trip diary into every prompt is how you blow the window and get amnesia.
Full Explanation
Developers must design context management including accumulation, retrieval, injection, and compaction for single- and multi-agent systems. Option B correctly combines structured state, selective retrieval/injection, and compaction while preserving critical entities—addressing sliding-window failures called out in multi-agent ops.
Option A causes context bloat and cost/latency explosions.
Option C destroys continuity needed for long-running cases.
Option D is not a scalable design and eventually still hits model limits.
Exam tip: Compaction + structured memory + selective injection is the expected pattern for long-running multi-agent context.