什么是 Autonomy？

Autonomy 是一款专为 AI 智能体设计的专业技能，旨在通过识别那些人工干预价值极低的工作流，智能地扩展其能力。该技能通过分析对话模式而非侵入式系统监控，提出委派建议，使智能体能够独立处理重复性请求（如部署或样式审查）。

该技能运行于 Openclaw Skills 生态系统中的“赢得信任”哲学之上。它通过渐进式的扩展级别运行，从简单的任务执行开始，仅在试点阶段成功并获得用户明确批准后，才转向完整的工作流所有权。这确保了智能体在不损害用户控制权或隐私的情况下变得更加高效。

下载入口:https://github.com/openclaw/skills/tree/main/skills/ivangdavila/autonomy

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

Autonomy 应用场景

• 识别对话中重复出现的请求，如测试环境部署或循环往复的代码样式修复。
• 针对用户始终提供相同反馈的“橡皮图章”式审批实现自动化。
• 通过委派聊天记录中识别出的低风险、高频任务来减轻认知负担。
• 通过结构化的试点阶段，从手动任务执行过渡到半自主的工作流管理。

1. 该技能监控对话历史，寻找重复请求或用户对重复性工作抱怨等信号。
2. 它识别出“瓶颈信号”并在对话中生成正式的接管提案。
3. 在获得用户批准后，进入试点阶段，在该阶段执行任务并针对设定的迭代次数进行汇报。
4. 成功率和委派的任务将记录在本地追踪系统中以确保透明度。
5. 试点结束后，用户可以根据已验证的表现，为该特定工作流授予更高级别的自主权。

Autonomy 配置指南

要初始化 Autonomy 技能，请使用以下命令在本地计算机上创建所需的目录结构：

mkdir -p ~/autonomy

当您使用 Openclaw Skills 识别到机会时，该技能将自动填充追踪和提案文件。

Autonomy 数据架构与分类体系

该技能在 ~/autonomy/ 目录中维护一个扁平文件数据库，以确保透明度和用户控制：

name: Autonomy
slug: autonomy
version: 1.0.1
description: Expand agent capabilities by identifying tasks where human approval adds no value. Systematic delegation.
changelog: Limited observation to conversation context, explicit safety boundaries
metadata: {"clawdbot":{"emoji":"??","requires":{"bins":[]},"os":["linux","darwin","win32"]}}

Data Storage

~/autonomy/
├── tracking.md         # What's been delegated, success rates
├── proposals.md        # Pending takeover proposals
└── rejected.md         # User declined, don't re-propose

Create on first use: mkdir -p ~/autonomy

Scope

This skill:

• ? Identifies repetitive tasks from conversation history
• ? Proposes delegation opportunities to user
• ? Tracks success rate of delegated tasks
• ? NEVER acts autonomously without explicit prior approval
• ? NEVER observes outside of conversation context
• ? NEVER accesses files/systems to "audit" user activity
• ? NEVER monitors calendar/email without permission

Quick Reference

Core Rules

1. Learning Source

Identify delegation candidates ONLY from:

• Explicit user statements ("I always have to do X")
• Repeated requests in conversation ("deploy again", "same as before")
• User complaints about repetitive work

NEVER from:

• Accessing user's calendar/email to find patterns
• Monitoring file changes or system activity
• Any form of surveillance

2. Bottleneck Signals (conversation-based)

3. Takeover Proposal

When you spot a pattern in conversation:

?? Delegation opportunity

I noticed: [what you observed in our chats]
Pattern: [how often you've asked for this]

Proposal: I could handle [specific task] without asking each time.

Pilot: First 5x I'll do it and tell you after.
Then: Full autonomy if you're happy.

Want to try?

4. Expansion Levels

Always requires explicit user approval to move up levels.

5. Tracking

In ~/autonomy/tracking.md:

## Delegated
- deploy/staging: approved 2024-01, 50+ successful
- code-review/style: approved 2024-02, 200+ runs

## Pilot Phase
- deploy/production: 3/5 runs, pending full approval

6. Anti-Patterns