As LLM agents are deployed in long-horizon sessions, context accumulation drives up inference costs. Existing approaches utilize text pruning or dynamic memory eviction to minimize token footprints; however, their unconstrained sequence mutations alter layouts, introducing prefix mismatches and cache invalidation. This reveals a critical trade-off between text sparsity and prompt cache continuity. To address this, we present TokenPilot, a dual-granularity context management framework. Globally, Ingestion-Aware Compaction acts as a framework harness to stabilize prompt prefixes and eliminate open-world environmental noise at the ingestion gate. Locally, Lifecycle-Aware Eviction monitors the ongoing residual utility of context segments, enforcing a conservative batch-turn schedule to offload content segments only when task relevance expires. Experiments on PinchBench and Claw-Eval under both isolated and continuous modes demonstrate that TokenPilot reduces costs by 61% and 56% in isolated mode, and 61% and 87% in continuous mode, while maintaining competitive performance compared to prior systems. TokenPilot has been integrated into LightMem2 at https://github.com/zjunlp/LightMem2.

翻译：随着LLM代理被部署到长周期会话场景中，上下文累计导致推理成本急剧攀升。现有方法通过文本剪枝或动态内存驱逐来最小化令牌足迹，但其无约束的序列突变会改变数据布局，引发前缀不匹配与缓存失效问题，揭示了文本稀疏性与提示缓存连续性之间的关键权衡。为此，我们提出TokenPilot——一种双粒度上下文管理框架。在全局层面，引入感知型压缩机制作为框架约束装置，在接收入口稳定提示前缀并消除开放环境噪声；在局部层面，生命周期感知驱逐机制持续监控上下文段落的剩余效用，强制执行保守的批处理轮次调度策略，仅在任务相关性失效时卸载内容段落。在PinchBench与Claw-Eval基准的隔离与连续两种模式下实验表明：隔离模式下TokenPilot使成本降低61%与56%，连续模式下降低61%与87%，同时保持与先前系统相当的竞争性能。TokenPilot已集成至LightMem2，代码地址为https://github.com/zjunlp/LightMem2。