dsh-block-ptc
mengwEX
AI-managed Auto Compact for DeepSeek Harness: the agent decides when to compact context and what to retain based on task progress, without waiting for a fixed token threshold.
gulagala00131Updated today
续聊时只发新内容和上一条回复的 ID,不再每次重传整段历史。没有界面,也不额外发探测请求。 Continue Responses API conversations in DeepSeek Harness by sending only new content and the previous response ID. No UI, no probe requests.
Practical score
78 / 100
Setup
Ready to use
License
MIT
Language
JavaScript
Copy the command, or hand the whole task to your harness agent.
git clone https://github.com/gulagala001/dsh-response-id-reuse.gitPlease install the DeepSeek Harness plugin "dsh-response-id-reuse" for me. - GitHub repository: gulagala001/dsh-response-id-reuse (https://github.com/gulagala001/dsh-response-id-reuse) - Type: a Cordis plugin (installed into a DSH profile) - The plugin declare…
A cleaned excerpt from the project's own documentation.
DSH Response ID Reuse 续聊时不再每次重传整段历史,只发新内容和上一条回复的 ID。 DeepSeek Harness 插件 简体中文 · English 安装 需要: DSH 0.2.2-rc.2 ,Node.js 24.9.0+ 。 在 DSH 里打开 插件 → + 添加插件 ,粘贴: 校验文件:…
Open on GitHubSimilar tools from the same corner of the ecosystem.
mengwEX
AI-managed Auto Compact for DeepSeek Harness: the agent decides when to compact context and what to retain based on task progress, without waiting for a fixed token threshold.
mrzturn
dsh-switchman who is a plugin for deepseek harness
yoza10635
Guarded context compaction for DeepSeek Harness (dsh): the LLM proposes, deterministic guards dispose — eager per-atom shrink (extract/summary/false under verbatim guards) + lazy reference-graph eviction (0-LLM) + byte-exact recall from an append-only log. 守卫式上下文压缩:LLM 只提议、确定性守卫裁决——逐原子缩放 + 惰性引用图剪枝(0-LLM)+ 追加式日志逐字节召回;压缩率精确兑现,历史永不销毁。