ZK-Andy/dsh-continual-evolve 预览 preview

ZK-Andy/dsh-continual-evolve

DeepSeek Harness 的持续自我进化插件:基于会话轨迹提炼出有版本管理、可审计、可回滚的 Harness 状态,并配以基准驱动的验证循环。

Project Overview项目介绍

This is a native DeepSeek Harness plugin built exclusively for the DSH ecosystem, adding continuous self-evolution capabilities to AI agents running on DSH. It provides a versioned, auditable, rollback-safe state layer that stores prompt notes, memories, skills, and subagent specifications refined from agent session trajectories. All core safety features, including schema validation, atomic writes, snapshots, versioning, audit trails, and acceptance decisions, are enforced entirely in code rather than relying on LLM prompt discipline. You can install it via two methods, either from npm by running dsh plugin add dsh-continual-evolve or directly from source with dsh plugin add ZK-Andy/dsh-continual-evolve.

It addresses the common problem of AI agents accumulating reusable experience during sessions that gets lost when the session ends, turning that scattered experience into structured, persistent state objects. It supports both session-local scope and cross-session global scope knowledge management, with built-in promotion guards that only allow high-quality, non-duplicate content into the global store. It also offers deterministic rollback, candidate refinement benchmarking, and automatic storage cleanup. This plugin is ideal for DSH users who need to build up long-term agent capabilities and continuously improve the quality of agent output.

This plugin requires Node.js version ^22.19 or >=24 to run, and you need to restart the dsh web service after installation or update for changes to take effect. It is released under the permissive MIT open source license, so it is completely free to use and modify. After installation, users can adjust dozens of configuration parameters via a config file, including automatic review interval, number of injected lines, and encryption keys. First-time users do not need to modify any settings, as the default configuration works out of the box for most use cases.

dsh-continual-evolve 是专为 DeepSeek Harness 开发的原生插件,为 DSH 提供 AI 代理持续自我进化能力。它是一个可版本化、可审计、支持安全回滚的代理状态层,可存储从会话轨迹中提炼的提示笔记、记忆、技能、子代理规范等内容。所有核心安全特性,包括模式验证、原子写入、快照、版本控制、审计追踪、接受决策,全部由代码强制执行,不依赖大语言模型的提示词规则。

它解决了 AI 代理在会话中积累可复用经验,但是跨会话容易丢失的问题,可将零散经验转化为结构化的状态对象。它支持会话本地范围、跨会话全局范围的知识管理,自带自动晋升防护机制,仅允许高质量、非重复知识进入全局存储。它还支持确定性回滚、候选优化基准测试、存储自动清理等功能,适合需要长期积累代理能力、持续优化代理输出质量的 DSH 用户。

该插件要求 Node.js 版本为 ^22.19 或 >=24,可通过 npm 或 GitHub 源码两种方式安装,安装或更新后需要重启 dsh web 服务即可生效。它采用 MIT 许可证开源,完全免费使用。用户安装后可通过配置文件调整自动审查间隔、注入行数、加密密钥等数十项参数,首次使用无需额外修改配置,默认参数即可正常运行。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 note1 项提示
  • 19 stars - an early-stage project星标 19,属于早期项目
DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

  • DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
  • External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时(Electron / Python / Docker 等)是否齐备
  • Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突:命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

  • Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致,是否归档或长期停更
  • Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
  • curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
  • Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
  • Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需(filesystem / network / shell / clipboard)
  • Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限,以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照,不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装(适合开发者)

CLI Install命令行安装

dsh plugin add dsh-continual-evolve

把 ZK-Andy/dsh-continual-evolve 加入你的 DSH 配置(web profile)即可启用。

READMEREADME

dsh-continual-evolve

中文 | English

awesome · DSH plugin npm CI License: MIT Node Tests Coverage · statements Coverage · branches Coverage · functions

Continual self-evolution for DeepSeek Harness: a versioned, auditable, rollback-safe harness state layer — prompt notes, memories, skills, subagent specs — refined from session trajectories.

The model proposes, the code guarantees. Every mechanical safety property — schema validation, atomic writes, snapshots, versioning, audit trail, acceptance decisions — is enforced in code, never by prompt discipline.

Why

Agents accumulate reusable experience (repeated failures, durable facts, reusable procedures) and forget it next session. This plugin turns that experience into first-class state:

  • Three scopes with merge semantics (global < project < local): local per-session staging, project per-workspace cross-session store, global cross-project — plus mechanical promotion guards so only portable, substantial, non-duplicate knowledge reaches global
  • Typed one-fact memories: every memory entry carries a recall type (user | feedback | project | reference); pitfalls (feedback) must include Why + How to apply
  • Moment-driven background memory agent: a bounded ZCode-style loop wakes only at low-frequency moments (compaction, goal-blocked streaks, the session-close drain, manual wrapup), searches the frozen memory manifest, and proposes memory-only edits through a closed tool set; it cannot call agents, MCP, the network, or write source files. Successful turns never cost an LLM call — conversation writes own routine sedimentation
  • Memory recall, projection, and receipts: evolve_recall reads back full memory content by query/kind/scope/type; every memory apply also materializes a readable MEMORY.md index plus one fact file per entry; each extraction lands a unified audit receipt (no-op/applied/declined with duration and turn stats) and only applied outcomes notify the session
  • Deterministic rollback: inverse edits generated from applied results — no LLM re-guessing
  • Benchmark loop: candidate refinements are evaluated against frozen cases by a separate scorer before acceptance (rubric encrypted at rest)
  • Store hygiene: /evolve consolidate turns write-time conflict hints and zero-use staleness into one approved, fully reversible batch of archives — with merge, near-duplicate content folds into the surviving original

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要,完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

← 上一个 Prev dsh-stock-market 下一个 Next dsh-hdc-bridge →