DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

Local Memory — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
L

dsh-local-memory

Local Memory

DSH local cross-session memory plugin: capture turn summaries, inject recent days, and search all historical memory with pure JS keyword matching.

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:caopu16/dsh-local-memory#20b1e4d24a6388febb8f23ad9fe50237dc9772a8
READMECompatibilityVersions

Compatibility and provenance

Local Memory is published as dsh-local-memory and currently resolves to version 0.1.0. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
8/27/2026

Versions

0.1.0stable
8/27/2026

Related plugins

Loading related plugins…

Latest
0.1.0
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
any
License
Not declared
Source
github
GitHub
★ 1
Weekly downloads
0
Last push
8/27/2026
View source ↗
README badge

Click the badge to copy Markdown for your README.

Do you maintain this Plugin?Claim benefit · Priority security scan

Verify the GitHub repository declared in package.json to manage this listing. After you claim it, Hub will prioritize a security scan of the current version and publish the result when it passes.

Claim this Plugin →
Report an issue

Related plugins

More verified plugins in memory-context.

Memsearch Dsh@zilliz/memsearch-dshMemSearch plugin for DeepSeek Harness: shared markdown memory across agents, with capture, pre-step context injection, memory-recall skill, and a skill-candidate review panel.Reme@agentscope-ai/remeReMe client and memory integrations for TypeScript agentsStratagate Dshstratagate-dshRecent conversations stay vivid. Older ones fade into summaries, not oblivion. StrataGate gives DeepSeek Harness six-layer, time-decaying memory, while lasting events and relationships settle into a knowledge graph. Bring your memories from other AIs withMeow Memorymeow-memoryCross-session project memory for DeepSeek Harness: seven-layer SQLite memory, first-turn snapshot injection, per-message keyword hits, memory_remember/search/project tools, automatic reflection with reflection-fold UI, and idle-triggered dream consolidati

README

dsh-local-memory

DSH(DeepSeek Harness)本地跨会话记忆插件:捕获每轮对话摘要、注入最近几天记忆,并提供 memory_search 工具按需检索全部历史记忆。

纯 JS / 纯文本实现,零 native 依赖。不依赖 memsearch CLI / pymilvus / numpy,因此不受本机 CPU 达不到 x86-64-v2 基线的影响。

功能

能力说明
capture每轮 turn/end 后把用户与助手对话摘要成 markdown bullet,追加到 <project>/.memsearch/memory/YYYY-MM-DD.md
inject每轮第 1 步自动注入最近 injectDays 天(默认 2 天)的记忆快照
memory_search纯 JS 关键词匹配,翻全部日文件(不止最近 2 天),供模型主动捞更早记忆
兼容 opencode与 opencode memsearch 插件共用 .memsearch/memory/ 目录和锚点格式,互读互不覆盖

安装

官方方式(推荐)

项目声明了 DSH bundle(dsh.bundle.patch),可以通过 DSH 官方插件命令安装:

# 已安装 dsh CLI 时
dsh plugin --profile web add github:caopu16/dsh-local-memory

# 从 deepseek-harness 源码运行时
cd /path/to/deepseek-harness
pnpm dsh plugin --profile web add github:caopu16/dsh-local-memory

lib/ 构建产物已提交,因此 git 源码安装不需要运行 prepare 构建脚本,也不会触发 pnpm ≥10 的 allowBuilds 拦截。

本地开发安装(备选)

cd ~/dsh-local-memory
bash scripts/install.sh

安装脚本会:

  1. 在项目 node_modules/@deepseek-ai/ 下创建 dsh-tools 和 dsh-llm 的本地软链(指向 ~/deepseek-harness 里的包),让独立项目也能解析 DSH 同款模块。
  2. 修改 ~/.dsh/profiles/web/cordis.patch.yml,把 local-memory 插件路径指向 ~/dsh-local-memory/src/index.mjs。

安装完成后重启 DSH 并新开一个会话,memory_search 工具才会出现在新会话里;已打开的会话不会自动获得新工具。

配置

官方安装后,插件由 bundle 自带的 cordis.patch.yml 挂载,行名为包名 dsh-local-memory。要覆盖配置,在 ~/.dsh/profiles/web/cordis.patch.yml 中 patch 同一行:

- id: local-memory
  name: 'dsh-local-memory'
  config:
    injectDays: 3
    searchMaxResults: 10
    provider: ''
    model: ''
配置项默认值说明
captureEnabledtrue是否捕获新对话
injectEnabledtrue是否注入最近记忆
searchEnabledtrue是否注册 memory_search 工具
summarizeEnabledtrue是否用 LLM 摘要;关掉则写原文
injectDays2注入时回看几个日文件
injectMaxLinesPerFile40每个日文件最多注入多少行
captureMaxChars24000单轮原文送摘要器的字符上限
summarizeMaxTokens1200摘要最大 token
summarizeTimeoutMs120000摘要超时
minPromptLength8用户消息短于此长度不捕获
searchMaxResults20memory_search 单次最多返回几个轮次块
searchMaxLinesPerResult12每个轮次块最多返回几行正文
provider / model空摘要路由;留空自动用当前 agent 的模型

使用

日常无需手动操作:capture 自动记录,inject 自动把最近记忆塞进上下文。

当话题可能早于注入窗口时,模型会主动调用:

memory_search(query: "flyway V22", top_k: 5)

返回结果按关键词/子串命中打分排序,包含日期文件、轮次标题和正文片段。

项目结构

~/dsh-local-memory/
├── src/
│   ├── index.mjs       # 插件入口:apply / name / inject
│   ├── constants.mjs   # 默认配置、marker、摘要系统提示
│   ├── util.mjs        # 日期、路径、内容块等公共工具
│   ├── capture.mjs     # turn 渲染、摘要、写日文件
│   ├── inject.mjs      # 最近记忆读取与快照消息
│   └── search.mjs      # 纯 JS 关键词召回与结果格式化
├── lib/                # 构建产物(已提交,供官方 dsh plugin 安装使用)
├── cordis.patch.yml    # DSH bundle 补丁层声明
├── scripts/
│   ├── build.mjs       # 把 src/ 构建到 lib/
│   ├── install.sh      # 本地开发安装到 DSH web profile
│   └── smoke-test.mjs  # 加载插件并真实执行 memory_search
├── test/
│   └── search.test.mjs # search 单元测试(node:test)
├── docs/
│   ├── design.md       # 设计说明
│   └── usage.md        # 详细使用与排障
├── package.json
└── README.md

测试

cd ~/dsh-local-memory
npm test          # 单元测试(不需要安装)
npm run smoke     # 冒烟测试(需要先执行 install.sh,且存在 jereh 记忆目录)

记忆文件格式

与 opencode memsearch 插件对齐:

# 2026-08-26

## Session 08:49

### 08:49
<!-- session:ses_xxx turn:msg_xxx db:/home/ubuntu/.local/share/opencode/opencode.db -->
- User 要求检查数据库与前端多语言 JSON 文件。
- Assistant 执行了 i18n 同步。