Configs2026年7月20日·1 分钟阅读

LifeOS — AI Intent Engineering Platform by Daniel Miessler

An open-source AI harness that conveys your goals and context to any AI on every task, created by the author of Fabric.

Agent 就绪

先审查再安装

这个资产需要先审查。复制的指令会要求 Agent dry-run、列出写入项,确认后再继续。

Needs Confirmation · 66/100策略:需确认
Agent 入口
任意 MCP/CLI Agent
类型
Skill
安装
Single
信任
信任等级:Established
入口
LifeOS
先审查命令
npx -y tokrepo@latest install 8ac79edc-8418-11f1-9bc6-00163e2b0d79 --target codex

先 dry-run,确认写入项后再运行此命令。

Introduction

LifeOS is an open-source intent engineering platform from Daniel Miessler, creator of Fabric. It bridges the gap between what you ultimately want and what your AI tools actually do by maintaining a persistent context of your goals, preferences, and projects across every interaction.

What LifeOS Does

  • Maintains a structured profile of your goals, preferences, and active projects
  • Injects relevant personal context into every AI interaction automatically
  • Provides task-specific workflows for coding, writing, business, and life management
  • Tracks progress toward goals and surfaces relevant actions
  • Integrates with multiple AI providers and coding agents

Architecture Overview

LifeOS is a TypeScript application that stores your intent graph locally in structured files. When you invoke any AI task, it reads your profile, selects relevant context based on the task type, and prepends it to the prompt. This ensures the AI always understands your broader objectives rather than treating each request in isolation.

Self-Hosting & Configuration

  • Clone the repository and install Node.js dependencies
  • Configure API keys for your preferred LLM providers in .env
  • Define your personal context in structured Markdown files under the context/ directory
  • Customize task templates in the workflows/ directory
  • Run locally or deploy as a lightweight web service

Key Features

  • Intent-first design ensures AI understands the why behind every task
  • Created by the author of Fabric, building on lessons from that project
  • Works across coding, writing, business strategy, and personal productivity
  • Modular workflow system lets you define custom task pipelines
  • Provider-agnostic architecture supports any OpenAI-compatible API

Comparison with Similar Tools

  • Fabric — Miessler's earlier project focused on patterns; LifeOS adds persistent personal context
  • Claude Code CLAUDE.md — project-scoped context only; LifeOS adds life-wide goal tracking
  • Mem0 — memory layer for AI apps; LifeOS focuses on intent and goals rather than conversation memory
  • Custom GPTs — locked to OpenAI; LifeOS is provider-agnostic and open source

FAQ

Q: How does LifeOS differ from Fabric? A: Fabric provides reusable AI patterns. LifeOS builds on that by adding a persistent personal context layer that automatically informs every AI interaction with your goals and preferences.

Q: Do I need coding experience to use it? A: Basic command-line familiarity is enough. Configuration is done through Markdown files and environment variables.

Q: Which AI providers work with LifeOS? A: Any provider with an OpenAI-compatible API, including OpenAI, Anthropic, local Ollama instances, and more.

Q: Is my personal data sent to third parties? A: Your context files stay local. Only the relevant portions are sent to your configured LLM provider as part of prompts.

Sources

讨论

登录后参与讨论。
还没有评论,来写第一条吧。

相关资产