latent-briefing

1# Latent Briefing and KV Cache Memory Sharing
2 
3Hierarchical multi-agent systems often pay for the same context twice. The orchestrator accumulates a long reasoning trajectory, but each worker usually receives only a narrow text handoff such as a subtask prompt plus raw document slices. Passing the full trajectory fixes coverage but drives token cost up on every worker call. Summarization introduces latency and information loss. Retrieval helps with document access but does not preserve the orchestrator's evolving reasoning state.
4 
5Latent Briefing addresses this by sharing memory at the **representation level** rather than the text level. The core idea is to compact the orchestrator trajectory in the worker model's KV cache, keeping positions that are most relevant to the **current worker task**. The method builds on **Attention Matching (AM)** KV cache compaction and adapts it for inference-time multi-agent handoff with task-guided queries, a shared token mask across heads, and robust thresholding.
6 
7## When to Activate
8 
9Activate this skill when:
10 
11- Designing orchestrator-worker or supervisor-specialist systems where workers need access to prior orchestrator state without replaying the full trajectory as text
12- Evaluating alternatives to LLM summarization or RAG for cross-agent state transfer
13- Implementing or studying **KV cache compaction** as a first-class inference primitive, not only prefix caching of identical prompts
14…