Automatic session retrospective: after a conversation wraps up, distills useful experience into reusable skills stored in the skill library for the model to call, with signal-accelerated triggers, manual runs, an approval mode, and skill-library governance (archive/restore, never deletes directly).
Install
# from npm (prebuilt)
dsh plugin --profile web add @weibaohui/hermes-loop
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:weibaohui/hermes-loop
Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).
README
This plugin publishes its README in Chinese only.
自动复盘插件：对话收尾后自动复盘，把有价值的经验蒸馏成可复用的技能（skill），把值得长期记住的事实与偏好写入跨会话记忆。
核心功能
- 自动复盘：对话达到阈值（默认 10 轮或 10 次工具调用）且过了冷却期，自动触发一次复盘；结论回显到来源会话
- 经验 → 技能：复盘产出的经验自动写成 skill 存入技能库，模型之后可以直接调用
- 信号加速触发：用户中断、工具连续失败、纠正词命中时跳过阈值提前复盘；纠正词可在面板自定义，命中率有统计
- 手动「立即复盘」：面板一键触发，运行过程实时可见
- 三种模式：自动写入 / 写入前需审批 / 仅记录不写入，设置面板即时切换
- 技能治理：技能使用统计、僵尸技能识别；Curator 自动维护技能库（活跃 → 归档 → 可恢复，永不直接删除）
- 跨会话记忆：复盘按需把环境/项目事实、约定教训写入
~/.dsh/memory/MEMORY.md，把用户画像与偏好写入USER.md（对齐 Hermes 的记忆分层）；记忆全文以会话首冻结方式注入每个对话，写入或手改后下个会话生效。按需产出——多数复盘不产生记忆，不为写而写 - 工作区记忆（v0.6，参考 Claude Code 的 per-project 记忆）：复盘自动判断归属——只适用当前项目的事实（端口、路径、项目特有约定）写入
~/.dsh/memory/workspaces/<工作区>/MEMORY.md，跨项目通用约定进全局库，用户画像恒全局；两层拼接注入，每层独立 2200 字符配额。无开关，常开
安装
dsh plugin --profile web add @weibaohui/hermes-loop -w
装完重启 dsh web 即生效。
使用
- 打开 Web UI → 设置页 → hermes-loop 面板
- 默认自动模式即可工作；想更稳妥可切换为「审批」模式（复盘写入前需你确认）
- 触发阈值、冷却时间、纠正词表均可在面板调整
- 复盘产出的技能进入技能库后，模型通过
skill工具直接使用；使用统计与归档管理也在面板里 - 沉淀记录、使用统计与 Curator 表里的技能名后都有「查看」按钮，弹窗直接看 SKILL.md 内容（Markdown 渲染；项目级技能也能看）
- 记忆条目在面板查看用量与内容；增删改直接编辑
~/.dsh/memory/MEMORY.md/USER.md（全局层）或~/.dsh/memory/workspaces/<工作区>/MEMORY.md（工作区层），一行一条、§开头，下个会话生效
联系我 :飞书群
版本兼容性
本插件与 DeepSeek Harness（@deepseek-ai/dsh）的版本对应关系：
| 插件版本 | 适配 dsh 版本 | 备注 |
|---|---|---|
| 0.1.17 | 0.2.0-rc.2 | 当前版本；复盘 prompt / 记忆快照 / 收尾沉淀段跟随 DSH 语言设置（locale.preference），不重启切换即生效；e2e 测试清理竞态修复 |
| 0.1.16 | 0.1.7-rc.2 | 修复降级路径 Config 导出 null——宿主 settings schema() 不排除 null，会炸 settings.describe() 拖垮整份设置文档（同 dsh-continue#4） |
| 0.1.13 | 0.1.7-rc.2 | 修复宿主将 volatile 字段物化为 {} 导致的设置毒化（saneConfigValues 清洗 + 移除 Config 兼容字符串字段） |
| 0.1.12 | 0.1.7-rc.2 | 适配 0.1.7 settings 模型（导出 volatile Config，ctx.settings.update 持久化），面板改动重启不再丢失 |
| 0.1.11 | 0.1.7-rc.2 | 面板改动只在本次运行内生效（0.1.7 移除了 ctx.settings.register），重启还原 |
发版约定：每次发布新版本时，请在上表追加一行，记录该插件版本实际验证所用的
@deepseek-ai/dsh版本。package.json的engines.dsh声明最低支持版本；本表记录实际验证版本，二者配合使用。
Links
More in this category
vectorize-io/hindsight#coding-agents★ 51706
Hindsight, agent memory that learns: long-term project memory with auto recall and retain, knowledge pages, deep reflection, and per-repo memory banks.
volcengine/OpenViking#examples/dsh-memory-plugin★ 40368
OpenViking memory and context bundle for DeepSeek Harness: pre-step auto-recall and profile injection, session capture, `viking://` URI guarding, and recall/write memory tools backed by an OpenViking server.
agentscope-ai/ReMe#dsh★ 3655
Connects DeepSeek Harness to ReMe's local-first, self-evolving personal knowledge base: automatically captures completed main-agent conversations as user-owned Markdown memory, searches conversations and source material through reme_search with BM25, optional embeddings, and wikilink expansion, and schedules daily memory consolidation.
zilliztech/memsearch#MemSearch★ 2763
Shared Markdown memory for DSH and other coding agents, with automatic capture, pre-step context injection, searchable recall, and memory-to-skill self-evolution through a review panel.
vshulcz/deja-vu#extensions/dsh★ 1254
Reads the session files thirty-three other coding agents on this machine already wrote — Claude Code, Codex, Cursor, VS Code Copilot Chat, opencode, OpenClaw, Hermes, Kimi, Cline, Zed and more — including sessions from before it was installed: six tools (deja_recall, deja_session, deja_blame, deja_fix, deja_how, deja_remember), a /deja command, and optional automatic recall added to the runtime context. Local BM25 index, no LLM, no embeddings, no network (dsh plugin --profile web add dsh-deja).
adoresever/graph-memory★ 655
Traceable, searchable cross-session memory for DeepSeek Harness — conversation knowledge as typed graph nodes (TASK/SKILL/EVENT) and typed edges.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.