PluginNeeds setup

dsh-argp

yoza10635152Updated 2d ago

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。

agent compaction context-compression cordis deepseek-harness dsh dsh-plugin llm 效率提升会话管理模型配置

Practical score

77 / 100

Setup

Needs setup

License

MIT

Language

TypeScript

Install

Quick start

Copy the command, or hand the whole task to your harness agent.

git clone https://github.com/yoza10635/dsh-argp.git

DSH versionDSH ≥ 0.2.0-rc.1Setup required

Install with your agent

Please install the DeepSeek Harness plugin "dsh-argp" for me. - GitHub repository: yoza10635/dsh-argp (https://github.com/yoza10635/dsh-argp) - Type: a Cordis plugin (installed into a DSH profile) - The plugin declares a DSH version requirement of ^0.2.0-rc.1.…

From the README

A cleaned excerpt from the project's own documentation.

English 中文 dsh-argp — 双引擎上下文压缩：逐原子守卫压缩 + 引用图确定性剪枝 dsh-argp 是 DeepSeek Harness（dsh）的第三方上下文压缩引擎：替换"让模型自由改写历史"的摘要式压缩—— LLM 只提方案、确定性守卫负责处置，被压缩 / 被剪枝的原文保留在 append-only 日志、可按 seq 召回（文本块逐字；工具调用参数为 JSON 语义等价重建），压缩目标按字符口径收敛（0-LLM 图规则；字符 ≠ provider 实际 token，估算误差见「已知限制」）。双引擎形态、安装即挂、无需手工配置；显式 peratom: false 可退回纯 0-LLM 图剪枝（见安装与挂载）。三个核心机制：…

Open on GitHub

Project details

Repository: yoza10635/dsh-argp
by: yoza10635
License: MIT
stars: 15
Forks: 2
Updated: Nov 9, 2026
Created: Aug 17, 2026
Language: TypeScript
Indexed from: topic

Related entries

Similar tools from the same corner of the ecosystem.

View all

Plugin

dsh-model-context-catalog

HOWILLMAKEIT

DeepSeek Harness 插件：一键匹配模型上下文窗口，展示容量来源，支持勾选批量配置、已有窗口管理和手动编辑。

cliproxyapicontext-windowcordis

30Ready to use

78/100

Plugin

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.

ai-agentscontext-managementdeepseek-harness

103Ready to use

88/100

Plugin

dsh-recall-plugin

limbo947

消息撤回插件：回到发送该消息时的状态 DSH Message Recall Plugin: Return to the state when the message was sent

cordisdeepseek-harnessdsh

55Ready to use

85/100