autonomous-agent-harness

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Autonomous Agent Harness

自主Agent套件

Turn Claude Code into a persistent, self-directing agent system using only native features and MCP servers.

仅使用原生功能和MCP服务器，将Claude Code转化为持久化、自导向的Agent系统。

When to Activate

激活场景

User wants an agent that runs continuously or on a schedule
Setting up automated workflows that trigger periodically
Building a personal AI assistant that remembers context across sessions
User says "run this every day", "check on this regularly", "keep monitoring"
Wants to replicate functionality from Hermes, AutoGPT, or similar autonomous agent frameworks
Needs computer use combined with scheduled execution

用户需要持续运行或按计划运行的Agent
搭建定期触发的自动化工作流
构建跨会话记住上下文的个人AI助手
用户提到「每天运行这个」、「定期检查这个」、「持续监控」
需要复现Hermes、AutoGPT或同类自主Agent框架的功能
需要结合定时执行的计算机操作能力

Architecture

架构

┌──────────────────────────────────────────────────────────────┐
│                    Claude Code Runtime                        │
│                                                              │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌─────────────┐ │
│  │  Crons   │  │ Dispatch │  │ Memory   │  │ Computer    │ │
│  │ Schedule │  │ Remote   │  │ Store    │  │ Use         │ │
│  │ Tasks    │  │ Agents   │  │          │  │             │ │
│  └────┬─────┘  └────┬─────┘  └────┬─────┘  └──────┬──────┘ │
│       │              │             │                │        │
│       ▼              ▼             ▼                ▼        │
│  ┌──────────────────────────────────────────────────────┐    │
│  │              ECC Skill + Agent Layer                  │    │
│  │                                                      │    │
│  │  skills/     agents/     commands/     hooks/        │    │
│  └──────────────────────────────────────────────────────┘    │
│       │              │             │                │        │
│       ▼              ▼             ▼                ▼        │
│  ┌──────────────────────────────────────────────────────┐    │
│  │              MCP Server Layer                        │    │
│  │                                                      │    │
│  │  memory    github    exa    supabase    browser-use  │    │
│  └──────────────────────────────────────────────────────┘    │
└──────────────────────────────────────────────────────────────┘

┌──────────────────────────────────────────────────────────────┐
│                    Claude Code Runtime                        │
│                                                              │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌─────────────┐ │
│  │  Crons   │  │ Dispatch │  │ Memory   │  │ Computer    │ │
│  │ Schedule │  │ Remote   │  │ Store    │  │ Use         │ │
│  │ Tasks    │  │ Agents   │  │          │  │             │ │
│  └────┬─────┘  └────┬─────┘  └────┬─────┘  └──────┬──────┘ │
│       │              │             │                │        │
│       ▼              ▼             ▼                ▼        │
│  ┌──────────────────────────────────────────────────────┐    │
│  │              ECC Skill + Agent Layer                  │    │
│  │                                                      │    │
│  │  skills/     agents/     commands/     hooks/        │    │
│  └──────────────────────────────────────────────────────┘    │
│       │              │             │                │        │
│       ▼              ▼             ▼                ▼        │
│  ┌──────────────────────────────────────────────────────┐    │
│  │              MCP Server Layer                        │    │
│  │                                                      │    │
│  │  memory    github    exa    supabase    browser-use  │    │
│  └──────────────────────────────────────────────────────┘    │
└──────────────────────────────────────────────────────────────┘

Core Components

核心组件

1. Persistent Memory

1. 持久化内存

Use Claude Code's built-in memory system enhanced with MCP memory server for structured data.

Built-in memory (

~/.claude/projects/*/memory/

User preferences, feedback, project context
Stored as markdown files with frontmatter
Automatically loaded at session start

MCP memory server (structured knowledge graph):

Entities, relations, observations
Queryable graph structure
Cross-session persistence

Memory patterns:

undefined

使用Claude Code内置的内存系统，配合MCP内存服务器增强结构化数据存储能力。

内置内存 (

~/.claude/projects/*/memory/

用户偏好、反馈、项目上下文
以带frontmatter的markdown文件存储
会话启动时自动加载

MCP内存服务器 (结构化知识图谱):

实体、关系、观测记录
可查询的图谱结构
跨会话持久化

内存使用模式:

undefined

Short-term: current session context

短期：当前会话上下文

Use TodoWrite for in-session task tracking

使用TodoWrite进行会话内任务跟踪

Medium-term: project memory files

中期：项目内存文件

Write to ~/.claude/projects/*/memory/ for cross-session recall

写入~/.claude/projects/*/memory/实现跨会话召回

Long-term: MCP knowledge graph

长期：MCP知识图谱

Use mcp__memory__create_entities for permanent structured data Use mcp__memory__create_relations for relationship mapping Use mcp__memory__add_observations for new facts about known entities

undefined

使用mcp__memory__create_entities存储永久结构化数据使用mcp__memory__create_relations进行关系映射使用mcp__memory__add_observations存储已知实体的新信息

undefined

2. Scheduled Operations (Crons)

2. 定时操作（Crons）

Use Claude Code's scheduled tasks to create recurring agent operations.

Setting up a cron:

undefined

使用Claude Code的定时任务创建重复的Agent运行任务。

创建cron:

undefined

Via MCP tool

通过MCP工具

mcp__scheduled-tasks__create_scheduled_task({ name: "daily-pr-review", schedule: "0 9 * * 1-5", # 9 AM weekdays prompt: "Review all open PRs in affaan-m/everything-claude-code. For each: check CI status, review changes, flag issues. Post summary to memory.", project_dir: "/path/to/repo" })

mcp__scheduled-tasks__create_scheduled_task({ name: "daily-pr-review", schedule: "0 9 * * 1-5", # 工作日上午9点 prompt: "Review all open PRs in affaan-m/everything-claude-code. For each: check CI status, review changes, flag issues. Post summary to memory.", project_dir: "/path/to/repo" })

Via claude -p (programmatic mode)

通过claude -p（编程模式）

echo "Review open PRs and summarize" | claude -p --project /path/to/repo


**Useful cron patterns:**

| Pattern | Schedule | Use Case |
|---------|----------|----------|
| Daily standup | `0 9 * * 1-5` | Review PRs, issues, deploy status |
| Weekly review | `0 10 * * 1` | Code quality metrics, test coverage |
| Hourly monitor | `0 * * * *` | Production health, error rate checks |
| Nightly build | `0 2 * * *` | Run full test suite, security scan |
| Pre-meeting | `*/30 * * * *` | Prepare context for upcoming meetings |

echo "Review open PRs and summarize" | claude -p --project /path/to/repo


**常用cron模式:**

| 模式 | 定时规则 | 适用场景 |
|---------|----------|----------|
| 每日站会 | `0 9 * * 1-5` | 审查PR、Issue、部署状态 |
| 每周复盘 | `0 10 * * 1` | 代码质量指标、测试覆盖率统计 |
| 每小时监控 | `0 * * * *` | 生产环境健康度、错误率检查 |
| 夜间构建 | `0 2 * * *` | 运行全量测试套件、安全扫描 |
| 会前准备 | `*/30 * * * *` | 为即将到来的会议准备上下文 |

3. Dispatch / Remote Agents

3. 调度/远程Agent

Trigger Claude Code agents remotely for event-driven workflows.

Dispatch patterns:

bash

undefined

远程触发Claude Code Agent，实现事件驱动的工作流。

调度模式:

bash

undefined

Trigger from CI/CD

从CI/CD触发

curl -X POST "https://api.anthropic.com/dispatch"
-H "Authorization: Bearer $ANTHROPIC_API_KEY"
-d '{"prompt": "Build failed on main. Diagnose and fix.", "project": "/repo"}'

Trigger from webhook

从webhook触发

GitHub webhook → dispatch → Claude agent → fix → PR

GitHub webhook → 调度 → Claude Agent → 修复 → PR

Trigger from another agent

从另一个Agent触发

claude -p "Analyze the output of the security scan and create issues for findings"

undefined

claude -p "Analyze the output of the security scan and create issues for findings"

undefined

4. Computer Use

4. 计算机操作

Leverage Claude's computer-use MCP for physical world interaction.

Capabilities:

Browser automation (navigate, click, fill forms, screenshot)
Desktop control (open apps, type, mouse control)
File system operations beyond CLI

Use cases within the harness:

Automated testing of web UIs
Form filling and data entry
Screenshot-based monitoring
Multi-app workflows

利用Claude的computer-use MCP实现物理世界交互能力。

能力范围:

浏览器自动化（导航、点击、填写表单、截图）
桌面控制（打开应用、输入、鼠标控制）
超出CLI范围的文件系统操作

套件内适用场景:

Web UI自动化测试
表单填写和数据录入
基于截图的监控
多应用工作流

5. Task Queue

5. 任务队列

Manage a persistent queue of tasks that survive session boundaries.

Implementation:

undefined

管理可跨会话存续的持久化任务队列。

实现方式:

undefined

Task persistence via memory

通过内存实现任务持久化

Write task queue to ~/.claude/projects/*/memory/task-queue.md

将任务队列写入~/.claude/projects/*/memory/task-queue.md

Task format

任务格式

name: task-queue type: project description: Persistent task queue for autonomous operation

Active Tasks

活跃任务

PR #123: Review and approve if CI green
Monitor deploy: check /health every 30 min for 2 hours
Research: Find 5 leads in AI tooling space

PR #123: CI通过则审查并批准
监控部署：2小时内每30分钟检查一次/health接口
调研：在AI工具领域找到5个潜在线索

Completed

已完成

Daily standup: reviewed 3 PRs, 2 issues

undefined

每日站会：审查3个PR、2个Issue

undefined

Replacing Hermes

替代Hermes方案

Hermes Component	ECC Equivalent	How
Gateway/Router	Claude Code dispatch + crons	Scheduled tasks trigger agent sessions
Memory System	Claude memory + MCP memory server	Built-in persistence + knowledge graph
Tool Registry	MCP servers	Dynamically loaded tool providers
Orchestration	ECC skills + agents	Skill definitions direct agent behavior
Computer Use	computer-use MCP	Native browser and desktop control
Context Manager	Session management + memory	ECC 2.0 session lifecycle
Task Queue	Memory-persisted task list	TodoWrite + memory files

Hermes组件	ECC等效方案	实现方式
网关/路由	Claude Code调度 + crons	定时任务触发Agent会话
内存系统	Claude内存 + MCP内存服务器	内置持久化 + 知识图谱
工具注册表	MCP服务器	动态加载的工具提供方
编排层	ECC技能 + Agent	技能定义指导Agent行为
计算机操作	computer-use MCP	原生浏览器和桌面控制
上下文管理器	会话管理 + 内存	ECC 2.0会话生命周期
任务队列	内存持久化任务列表	TodoWrite + 内存文件

Setup Guide

搭建指南

Step 1: Configure MCP Servers

步骤1：配置MCP服务器

Ensure these are in

~/.claude.json

json

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["-y", "@anthropic/memory-mcp-server"]
    },
    "scheduled-tasks": {
      "command": "npx",
      "args": ["-y", "@anthropic/scheduled-tasks-mcp-server"]
    },
    "computer-use": {
      "command": "npx",
      "args": ["-y", "@anthropic/computer-use-mcp-server"]
    }
  }
}

确保

~/.claude.json

中包含以下配置：

json

{
  "mcpServers": {
    "memory": {
      "command": "npx",
      "args": ["-y", "@anthropic/memory-mcp-server"]
    },
    "scheduled-tasks": {
      "command": "npx",
      "args": ["-y", "@anthropic/scheduled-tasks-mcp-server"]
    },
    "computer-use": {
      "command": "npx",
      "args": ["-y", "@anthropic/computer-use-mcp-server"]
    }
  }
}

Step 2: Create Base Crons

步骤2：创建基础Crons

bash

undefined

bash

undefined

Daily morning briefing

每日早间简报

claude -p "Create a scheduled task: every weekday at 9am, review my GitHub notifications, open PRs, and calendar. Write a morning briefing to memory."

Continuous learning

持续学习

claude -p "Create a scheduled task: every Sunday at 8pm, extract patterns from this week's sessions and update the learned skills."

undefined

claude -p "Create a scheduled task: every Sunday at 8pm, extract patterns from this week's sessions and update the learned skills."

undefined

Step 3: Initialize Memory Graph

步骤3：初始化内存图谱

bash

undefined

bash

undefined

Bootstrap your identity and context

初始化你的身份和上下文

claude -p "Create memory entities for: me (user profile), my projects, my key contacts. Add observations about current priorities."

undefined

claude -p "Create memory entities for: me (user profile), my projects, my key contacts. Add observations about current priorities."

undefined

Step 4: Enable Computer Use (Optional)

步骤4：启用计算机操作（可选）

Grant computer-use MCP the necessary permissions for browser and desktop control.

为computer-use MCP授予浏览器和桌面控制所需的必要权限。

Example Workflows

示例工作流

Autonomous PR Reviewer

自主PR审查员

Cron: every 30 min during work hours
1. Check for new PRs on watched repos
2. For each new PR:
   - Pull branch locally
   - Run tests
   - Review changes with code-reviewer agent
   - Post review comments via GitHub MCP
3. Update memory with review status

Cron：工作时间每30分钟运行一次
1. 检查关注仓库的新PR
2. 对每个新PR：
   - 拉取分支到本地
   - 运行测试
   - 使用代码审查Agent审查变更
   - 通过GitHub MCP发布审查评论
3. 更新内存中的审查状态

Personal Research Agent

个人调研Agent

Cron: daily at 6 AM
1. Check saved search queries in memory
2. Run Exa searches for each query
3. Summarize new findings
4. Compare against yesterday's results
5. Write digest to memory
6. Flag high-priority items for morning review

Cron：每天早上6点运行
1. 检查内存中保存的搜索查询
2. 为每个查询运行Exa搜索
3. 汇总新发现
4. 和昨日结果对比
5. 将摘要写入内存
6. 标记高优先级项供早间审查

Meeting Prep Agent

会议准备Agent

Trigger: 30 min before each calendar event
1. Read calendar event details
2. Search memory for context on attendees
3. Pull recent email/Slack threads with attendees
4. Prepare talking points and agenda suggestions
5. Write prep doc to memory

触发时机：每个日历事件前30分钟
1. 读取日历事件详情
2. 搜索内存中与会者相关上下文
3. 拉取近期和与会者的邮件/Slack对话
4. 准备谈话要点和议程建议
5. 将准备文档写入内存

Constraints

约束说明

Cron tasks run in isolated sessions — they don't share context with interactive sessions unless through memory.
Computer use requires explicit permission grants. Don't assume access.
Remote dispatch may have rate limits. Design crons with appropriate intervals.
Memory files should be kept concise. Archive old data rather than letting files grow unbounded.
Always verify that scheduled tasks completed successfully. Add error handling to cron prompts.

Cron任务在隔离会话中运行——除非通过内存，否则不会和交互会话共享上下文
计算机操作需要显式权限授予，不要默认拥有访问权限
远程调度可能存在速率限制，设计cron时请设置合理的间隔
内存文件应保持精简，归档旧数据避免文件无限制增长
始终验证定时任务是否成功完成，在cron提示中添加错误处理逻辑