deepseekplugins

DeepSeek Harness plugin

dsh-agent-memory-jonah791

Agent-driven long-term memory for DeepSeek Harness: scoped memory (global + per-workspace), layered entries (fact/knowledge/episodic), time-bucket compaction (day→week→month→year), agent-decided

Workflow & Automation
Jump to install

Source facts

Repository
jonah791/dsh-agent-memory
Latest update
Aug 17, 2026
Category
Workflow & Automation
GitHub stars
2
Format
plugin
Catalog evidence
Upstream dsh.bundle evidence
Evidence path
package.json#dsh.bundle
Checked against
0.1.0-rc.8
Upstream check date
2026-08-20

This evidence comes from the upstream catalog. This site has not installed, run, or security-reviewed the plugin.

Install

Start with a prompt that asks an agent to review the GitHub repository and source. Switch to the command if you want to install it yourself.

Copy this prompt into DSH, Codex, or another agent and ask it to review the GitHub repository and source first.

Do not install or run any commands yet. Read this plugin's GitHub repository, README, and relevant source code. Then answer the questions below clearly and directly so I can decide whether it fits my needs:

1. What is this plugin, and what problem does it solve?
2. Who is it for, and what are its typical use cases?
3. How is it used after installation? Include one minimal example.
4. What known limitations or privacy, security, compatibility, or maintenance risks does it have?
5. Give a clear recommendation: recommend, conditionally recommend, or do not recommend, with reasons.

Distinguish statements documented by the repository, inferences from source code, and unknowns. If evidence is insufficient, say so explicitly. Do not guess or simply repeat the README.

GitHub: https://github.com/jonah791/dsh-agent-memory
Plugin: dsh-agent-memory-jonah791
Author: jonah791

Check the source files

Read the README and other files from this plugin directory before installing.

File explorer3 files
README.mdSource · read only

dsh-agent-memory — DSH 的 Agent 驱动长期记忆插件

为 DeepSeek Harness (DSH) 打造的智能体驱动长期记忆插件。 跨会话的经历库：主人偏好、项目知识、决策理由、学习沉淀——结构化写入、可检索读取、可整理压缩。

> 状态：v0.3（智能体核心联动已落地）。DSH 为预览版（0.1.0-rc），无兼容承诺。

安装

cd <你的 self-plugins 目录>
git clone https://github.com/jonah791/dsh-agent-memory.git
cd dsh-agent-memory
pnpm install
pnpm build

设计总纲

程序系统只是工具和框架，重要的决策与行为由智能体自己决定。

  • 框架层负责可靠的条件：存储持久化、作用域路由、时间刻度压缩、检索管道、去重与完整性、工具面形状
  • 智能体负责一切内容判断：记什么、怎么组织、何时 recall、更新与遗忘
  • 「爱丽丝为核心」分工总则：机制只保证不丢、知道、兜底，不替 agent 做内容决策

功能

分层记忆（L1 事实 / L2 知识 / L3 情景）

  • L1 fact：主人偏好、环境事实——key 精确覆盖更新
  • L2 knowledge：可复用知识、学习沉淀——同标题自动合并
  • L3 episodic：有结果的情景、经历时间线——时间桶压缩（日→周→月→年）
  • summary：压缩产物，冷归档保留（可深挖），索引高密度

作用域

  • global：全局记忆（主人偏好/环境），永远附加检索
  • workspace 级：按项目 .dsh/memory.yml 配置路由（层级、预算、时间刻度）

工具面（agent 自主调用）

工具用途
remember写入/覆盖/合并；返回 id + action
recall检索：关键词/层级/标签/时间过滤，相关度+新鲜度排序
memory_browse时间金字塔浏览（不知道有什么时的发现路径）
update / forget修订 / 归档（不进活跃检索，可深挖）
memory_stats各层/时间桶/归档计数

压缩即记忆（通道 C，与 dsh-agent-compact 联动）

会话压缩完成 → checkpoint 原文保底存档（episodic）+ inbox 通知（wakeup=false 排队不唤醒）——提炼与否、如何组织由 agent 自主决策，理由记入 source.reason。压缩在进程内已完整，不再写哨兵重启。

组合

- insert:
    - id: agent-memory
      name: dsh-agent-memory

依赖官方 storage 栈（storage / storage-json / storage-domain，web-app bundle 已提供）。

测试

node node_modules/typescript/lib/tsc.js -p tsconfig.json && node --test

165+ tests / 0 fail。

设计文档

  • [DESIGN.md](DESIGN.md) — 设计意图（输入通道、检索管道、防失控机制）
  • [IMPLEMENTATION.md](IMPLEMENTATION.md) — 工程契约（接口、验收、取证记录）

关联项目

相关

License

MIT

Back to searchOpen GitHub repository

More related plugins

View category