dsh-context Tool Visualises Agent Context Usage for DSH
Context cost in long-running agents usually hides inside tool results and sub-agent calls, and dsh-context makes those allocations traceable instead of guessing from a total token count. Builders running multi-agent workflows on DSH can use it to diagnose cache misses, runaway context growth and per-sub-agent spend, then rebalance prompts or trimming before costs climb.
A new tool called dsh-context provides a visual dashboard for tracking what fills an agent's context window in DSH. It breaks token usage down by system prompts, tool definitions, conversation messages and tool results to locate context bloat. It also logs context changes per turn or step, showing when compression or trimming happens. A Fleet view displays the relationship between main agents and sub-agents, their context occupancy, execution status and token costs.
一个让你看清Agent上下文窗口里到底装了什么、怎么涨起来的、钱花在哪儿的工具：dsh-context
专门给DSH用的，把原来不透明的信息，给你变成可视化、能追踪分析的仪表盘
帮你诊断为什么上下文会爆、成本会涨、缓存命中率低、agent/子agent到底在吃什么token
它有几个能力： 1、看清楚Agent的上下文为什么会膨胀 查看系统提示词、工具定义、对话消息、工具返回分别占用了多少Token，快速定位到上下文膨胀的来源
2、追踪上下文压缩前后发生了什么 按轮次/步骤记录上下文变化，观察Token如何增长、什么时候发生压缩和裁剪，以及前后发生了什么变化
3、观察多Agent的上下文和成本 查看上下文中的消息块、工具调用与工具结果构成，不是只给一个总Token数。通过Fleet视图查看主Agent与子Agent的关系、上下文占用、执行状态及Token成本
随着Agent开始能执行长时复杂任务，上下文有个专门监控诊断的工具确实很有必要
#dshcontext #AI上下文诊断工具 #DSH
