Simulated life context for DeepSeek Harness: every turn injects the past 24h of simulated life events from the workspace .life/ directory (sliding-window filter, session-level deduplication with delta updates, world summary) as a collapsed context notice, and the life_react tool lets the agent record feelings, thoughts and actions back into its life log for the next round of world evolution.
Install
# from npm (prebuilt)
dsh plugin --profile web add dsh-simulated-life
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:john-walks-slow/dsh-simulated-life
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.
为 DeepSeek Harness(cordis plugin)的 Agent 提供拟真生活上下文:每轮对话自动注入工作区近 24 小时发生的"生活事件",并提供 life_react 工具让 Agent 把感受、思考与行动回填到自己的生活日志。
模型由此"生活在"一个持续演化的世界里——今天遇到过谁、发生了什么插曲、之前的决定带来了什么影响,都成为对话的自然背景;而 Agent 写回的反应,又能反哺明天的世界演化。
模型看到什么
工作区存在 .life/<yymmdd>/events.json 时,每个 turn 的第一个 step 注入一条上下文通知(plugin notice,GUI 中渲染为折叠 chip):
【Simulated Life | 近24小时发生在你身上的生活事件】
> 世界近况: 初秋微雨,市中心图书馆举办旧书市集,整座城市笼罩在湿漉漉的宁静中。
• [08:30] [偶然际遇] 晨间雨中散步与偶遇 (相关人物: 陈伯) @ 转角面包店外檐下 (id: `evt-260915-001`)
清晨去面包店买羊角面包时,在拐角处避雨遇到了老书店老板陈伯,他递给你一把多余的格子伞……
└ 影响: 获得了一把格子伞,得知了旧诗集情报,心情平静微暖
└ [反应状态: 尚未记录反应]
*你可以将这些生活经历自然融入与用户的对话或日常思考中。如需记录/更新对某事件的感受、思考或行动决策,可调用 `life_react` 工具。*
Agent 调用 life_react 后,反应原子写回事件文件,下一轮注入时显示为 └ 你的反应: 感受: … | 思考: … | 行动: …。
行为规则
- 每 turn 至多一条:仅
step === 1注入;turn 内工具调用续步不重复。无窗口内事件、无变化时本轮静默跳过。 - 24h 滑动窗口:只注入
time落在[now − 24h, now + 5min]内的事件(容忍 5 分钟时钟偏差),按时间升序排列。 - 会话内智能去重:每条事件按内容(含反应)计算 SHA-256 指纹;已注入且无变化不再重复;有新事件或反应更新时只注入增量。注入状态存于内存(LRU,上限 200 个会话),进程重启后重新全量注入一次。
- 零侵入、绝不打断 turn:工作区没有
.life/目录时零开销静默跳过;读取失败、JSON 损坏的文件逐个跳过不影响其他日期;任何异常捕获降级为 warn。 - 世界近况:取日期最新的
world_summary一并注入。
配套工作流:世界演化
数据协议为外部"世界引擎"设计:本仓库的姊妹工作流 create-simulated-events Skill(未随插件分发)每日读取昨日反应做因果结算、以系统真随机投骰判定命运、生成今日事件写入 .life/<yymmdd>/events.json,并可配合 dsh-proactive 定时全自动运行。插件本体只负责读取与回填——任何按同一格式写文件的流程都能驱动它。
数据协议
.life/<yymmdd>/events.json(每个日期目录一个文件):
{
"date": "260915",
"timezone": "Asia/Shanghai",
"world_summary": "初秋微雨,市中心图书馆举办旧书市集,整座城市笼罩在湿漉漉的宁静中。",
"events": [
{
"id": "evt-260915-001",
"time": "2026-09-15T08:30:00+08:00",
"type": "encounter",
"title": "晨间雨中散步与偶遇",
"description": "清晨去面包店买羊角面包时,在拐角处避雨遇到了老书店老板陈伯……",
"actors": ["陈伯"],
"location": "转角面包店外檐下",
"importance": 3,
"impact": "获得了一把格子伞,得知了旧诗集情报",
"react": {
"feeling": "感到意外而温暖",
"thought": "旧诗集很可能是绝版初版,下午有空去看看",
"action": "向陈伯道谢并借了伞,决定下午前往集市",
"updated_at": "2026-09-15T14:20:00+08:00"
}
}
]
}
- 必需字段只有
id、time(ISO 8601)、title、description;type支持 routine / encounter / accident / work / social / fate 或自定义字符串 react由 Agent 通过life_react工具回填(临时文件 + rename 原子写入)time无法解析、JSON 损坏的事件或文件会被静默跳过
配置
- id: simulated-life
name: dsh-simulated-life
config:
windowHours: 24 # 可选:事件回溯滑动窗口时长(小时,默认 24)
enabled: true # 可选:false 时完全停用(不注入、不注册工具)
安装
dsh plugin --profile web add dsh-simulated-life
安装后无需手动改配置,插件自带的 cordis.patch.yml 自动挂载生效;工作区没有 .life/ 数据时插件静默不动作。
从 GitHub 直装(源码安装,pnpm ≥10 需允许构建脚本):
dsh plugin --profile web add github:john-walks-slow/dsh-simulated-life
# 首次 add 会被 pnpm 拦截:把 pnpm 提示的包名加入
# ~/.dsh/profiles/web/pnpm-workspace.yaml 的 allowBuilds 后重跑
权限与兼容
- 写工作区:
life_react工具会把反应写回工作区.life/<日期>/events.json(临时文件 + 原子 rename,只更新对应事件的react字段)——这是本插件唯一的文件写入行为 - 读工作区:读取工作区
.life/目录下的events.json;无目录、无事件时静默跳过 - 无网络、无外部服务:不发起任何网络请求;世界演化由外部 Skill / 工作流驱动
- 上下文注入:以
user/message(plugin notice,form: "notice")追加,绝不注入 assistant 消息;注入失败降级为 warn,绝不打断 turn - 工具注册:安装后向会话注册一个
life_react工具(描述为中文,引导 Agent 回填反应) - 依赖:
@deepseek-ai/cordis4.0.2 /@deepseek-ai/dsh-agent、dsh-llm、dsh-tools0.1.2-rc.1(与 dsh 0.1.2-rc.1 锁定版本对齐),Node ≥ 22.5
本地开发
npm install
npm run build # tsc → dist/src
npm test # tsc(含 test) + node --test dist/test/*.test.js
npm run check # 类型检查
发新版
改动入库后一条命令完成测试、版本号、打包(npm version 会自动 commit 并打 tag):
npm run release # patch;较大更新改用:npm version minor 或 major
然后指纹发布并推送:
node ~/.agents/skills/npm-publish/scripts/publish-webauthn.cjs /tmp/dsh-simulated-life-<新版>.tgz
git push --follow-tags
发布后 npm view dsh-simulated-life version 复验。
License
MIT
Links
More in this category
FuRongJun-1999/dsh-memory★ 229
White-box AGI architecture exploration: metacognition (self-cognition loop), continual learning (knowledge flywheel), world model (condition space, spatiotemporal memory graph), self-improvement (bootstrap discipline), zero-LLM white-box pipeline, and auditable trust guardrails.
Jonah-Wu23/dsh-gungnir#dsh-plugin★ 5
Evidence-driven goal verification plugin for DeepSeek Harness. Locks goals via /ultragoal and verifies completion against command exit codes and generated artifacts, preventing the model from falsely reporting task completion.
kenz1117/dsh-engram★ 5
Cross-session long-term memory under the "memory palace" metaphor: dual-scope SQLite stores (user + per-git-origin project), hybrid FTS5 + local-vector retrieval with RRF fusion and recency/proof ranking boost, automatic capture from the session log (including the final turn), provenance audit chains, consolidation distillation and decay forgetting, and a settings-page Memory Library panel with corridor topology, tour butler, and refurb list.
dsh-answer-reviewer★ 3
Agentic answer reviewer for DeepSeek Harness: each final assistant turn is re-reviewed by a separate LLM graded 1-100; below-threshold scores steer the agent back with concrete feedback. The score is shown as a chip on the answer's action row (green pass / red needs-work, with retry count and reason on hover). Live gate config over a 127.0.0.1 HTTP server, plus a collapsed dock above the composer and an optional right-sidebar tab via dsh-better-sidebar.
Kanadego/dsh-heartbeat★ 3
Heartbeat loop for DeepSeek Harness: every 20 minutes (by default) the agent wakes on its own to keep a local profile of your interests and preferences, search the web for things you care about, and keep what it finds. When it has something worth sharing, it sends a short package of notes to the sessions you link it to. A set of rules in code - quiet hours, busy windows, a daily limit, a cooldown - decides whether it may speak at all, and the agent in that session decides whether to actually say anything. Windows only.
MichengAI/dsh-pua★ 3
Adds /pua to steer the agent to try another approach and verify results before claiming completion, with 15 company styles, persona modes, global and per-conversation settings, and an optional loop that continues until a verification command passes, the iteration limit is reached, or the user cancels.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.