self-improvement

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Self-Improvement Skill

自我改进Skill

Log learnings and errors to markdown files for continuous improvement. Coding agents can later process these into fixes, and important learnings get promoted to project memory.

将学习内容和错误记录到Markdown文件中，以实现持续改进。编码Agent后续可将这些内容转化为修复方案，重要的学习内容会被提升至项目记忆库。

Quick Reference

快速参考

Situation	Action
Command/operation fails	Log to `.learnings/ERRORS.md`
User corrects you	Log to `.learnings/LEARNINGS.md` with category `correction`
User wants missing feature	Log to `.learnings/FEATURE_REQUESTS.md`
API/external tool fails	Log to `.learnings/ERRORS.md` with integration details
Knowledge was outdated	Log to `.learnings/LEARNINGS.md` with category `knowledge_gap`
Found better approach	Log to `.learnings/LEARNINGS.md` with category `best_practice`
Similar to existing entry	Link with `See Also` , consider priority bump
Broadly applicable learning	Promote to `CLAUDE.md` , `AGENTS.md` , and/or `.github/copilot-instructions.md`
Workflow improvements	Promote to `AGENTS.md` (OpenClaw workspace)
Tool gotchas	Promote to `TOOLS.md` (OpenClaw workspace)
Behavioral patterns	Promote to `SOUL.md` (OpenClaw workspace)

场景	操作
命令/操作失败	记录至 `.learnings/ERRORS.md`
用户纠正你	记录至 `.learnings/LEARNINGS.md` ，分类标记为 `correction`
用户需要缺失的功能	记录至 `.learnings/FEATURE_REQUESTS.md`
API/外部工具故障	记录至 `.learnings/ERRORS.md` ，并添加集成详情
知识过时	记录至 `.learnings/LEARNINGS.md` ，分类标记为 `knowledge_gap`
发现更优方法	记录至 `.learnings/LEARNINGS.md` ，分类标记为 `best_practice`
与现有条目相似	添加 `See Also` 链接，考虑提升优先级
广泛适用的学习内容	提升至 `CLAUDE.md` 、 `AGENTS.md` 和/或 `.github/copilot-instructions.md`
工作流改进	提升至 `AGENTS.md` （OpenClaw工作区）
工具注意事项	提升至 `TOOLS.md` （OpenClaw工作区）
行为模式	提升至 `SOUL.md` （OpenClaw工作区）

OpenClaw Setup (Recommended)

OpenClaw配置（推荐）

OpenClaw is the primary platform for this skill. It uses workspace-based prompt injection with automatic skill loading.

OpenClaw是使用此Skill的主要平台，它基于工作区的提示注入功能，可自动加载Skill。

Installation

安装

Via ClawdHub (recommended):

bash

clawdhub install self-improving-agent

Manual:

bash

git clone https://github.com/peterskoett/self-improving-agent.git ~/.openclaw/skills/self-improving-agent

通过ClawdHub安装（推荐）：

bash

clawdhub install self-improving-agent

手动安装：

bash

git clone https://github.com/peterskoett/self-improving-agent.git ~/.openclaw/skills/self-improving-agent

Workspace Structure

工作区结构

OpenClaw injects these files into every session:

~/.openclaw/workspace/
├── AGENTS.md          # Multi-agent workflows, delegation patterns
├── SOUL.md            # Behavioral guidelines, personality, principles
├── TOOLS.md           # Tool capabilities, integration gotchas
├── MEMORY.md          # Long-term memory (main session only)
├── memory/            # Daily memory files
│   └── YYYY-MM-DD.md
└── .learnings/        # This skill's log files
    ├── LEARNINGS.md
    ├── ERRORS.md
    └── FEATURE_REQUESTS.md

OpenClaw会将以下文件注入到每个会话中：

~/.openclaw/workspace/
├── AGENTS.md          # 多Agent工作流、委托模式
├── SOUL.md            # 行为准则、个性、原则
├── TOOLS.md           # 工具能力、集成注意事项
├── MEMORY.md          # 长期记忆（仅主会话）
├── memory/            # 每日记忆文件
│   └── YYYY-MM-DD.md
└── .learnings/        # 此Skill的日志文件
    ├── LEARNINGS.md
    ├── ERRORS.md
    └── FEATURE_REQUESTS.md

Create Learning Files

创建学习文件

bash

mkdir -p ~/.openclaw/workspace/.learnings

Then create the log files (or copy from

assets/

```
LEARNINGS.md
```
— corrections, knowledge gaps, best practices
```
ERRORS.md
```
— command failures, exceptions
```
FEATURE_REQUESTS.md
```
— user-requested capabilities

bash

mkdir -p ~/.openclaw/workspace/.learnings

然后创建日志文件（或从

assets/

复制）：

```
LEARNINGS.md
```
— 修正内容、知识缺口、最佳实践
```
ERRORS.md
```
— 命令失败、异常
```
FEATURE_REQUESTS.md
```
— 用户请求的功能

Promotion Targets

提升目标

When learnings prove broadly applicable, promote them to workspace files:

Learning Type	Promote To	Example
Behavioral patterns	`SOUL.md`	"Be concise, avoid disclaimers"
Workflow improvements	`AGENTS.md`	"Spawn sub-agents for long tasks"
Tool gotchas	`TOOLS.md`	"Git push needs auth configured first"

当学习内容被证明具有广泛适用性时，将其提升至工作区文件：

学习类型	提升至	示例
行为模式	`SOUL.md`	"保持简洁，避免免责声明"
工作流改进	`AGENTS.md`	"为长任务生成子Agent"
工具注意事项	`TOOLS.md`	"Git push需要先配置认证"

Inter-Session Communication

跨会话通信

OpenClaw provides tools to share learnings across sessions:

sessions_list — View active/recent sessions
sessions_history — Read another session's transcript
sessions_send — Send a learning to another session
sessions_spawn — Spawn a sub-agent for background work

OpenClaw提供工具以在会话间共享学习内容：

sessions_list — 查看活跃/近期会话
sessions_history — 读取另一个会话的记录
sessions_send — 向另一个会话发送学习内容
sessions_spawn — 生成子Agent处理后台工作

Optional: Enable Hook

可选：启用钩子

For automatic reminders at session start:

bash

undefined

如需在会话开始时自动提醒：

bash

undefined

Copy hook to OpenClaw hooks directory

复制钩子到OpenClaw钩子目录

cp -r hooks/openclaw ~/.openclaw/hooks/self-improvement

Enable it

启用钩子

openclaw hooks enable self-improvement


See `references/openclaw-integration.md` for complete details.

---

openclaw hooks enable self-improvement


完整详情请参见`references/openclaw-integration.md`。

---

Generic Setup (Other Agents)

通用配置（其他Agent）

For Claude Code, Codex, Copilot, or other agents, create

.learnings/

in your project:

bash

mkdir -p .learnings

Copy templates from

assets/

or create files with headers.

对于Claude Code、Codex、Copilot或其他Agent，在你的项目中创建

.learnings/

bash

mkdir -p .learnings

从

assets/

复制模板，或创建带标题的文件。

Logging Format

记录格式

Learning Entry

学习条目

Append to

.learnings/LEARNINGS.md

markdown

undefined

追加至

.learnings/LEARNINGS.md

：

markdown

undefined

[LRN-YYYYMMDD-XXX] category

[LRN-YYYYMMDD-XXX] 分类

Summary

摘要

One-line description of what was learned

一行描述所学内容

Details

详情

Full context: what happened, what was wrong, what's correct

完整上下文：发生了什么、哪里出错、正确做法是什么

Suggested Action

建议操作

Specific fix or improvement to make

具体的修复或改进方案

Metadata

元数据

Source: conversation | error | user_feedback
Related Files: path/to/file.ext
Tags: tag1, tag2
See Also: LRN-20250110-001 (if related to existing entry)

undefined

来源: conversation | error | user_feedback
相关文件: path/to/file.ext
标签: tag1, tag2
另请参阅: LRN-20250110-001（如果与现有条目相关）

undefined

Error Entry

错误条目

Append to

.learnings/ERRORS.md

markdown

undefined

追加至

.learnings/ERRORS.md

：

markdown

undefined

[ERR-YYYYMMDD-XXX] skill_or_command_name

Summary

摘要

Brief description of what failed

故障的简要描述

Error

错误信息

Actual error message or output

实际错误消息或输出

Context

上下文

Command/operation attempted
Input or parameters used
Environment details if relevant

尝试执行的命令/操作
使用的输入或参数
相关的环境详情（如有）

Suggested Fix

建议修复方案

If identifiable, what might resolve this

如果可识别，可能的解决方法

Metadata

元数据

Reproducible: yes | no | unknown
Related Files: path/to/file.ext
See Also: ERR-20250110-001 (if recurring)

undefined

可复现: yes | no | unknown
相关文件: path/to/file.ext
另请参阅: ERR-20250110-001（如果重复出现）

undefined

Feature Request Entry

功能请求条目

Append to

.learnings/FEATURE_REQUESTS.md

markdown

undefined

追加至

.learnings/FEATURE_REQUESTS.md

：

markdown

undefined

[FEAT-YYYYMMDD-XXX] capability_name

Requested Capability

请求的功能

What the user wanted to do

用户想要实现的操作

User Context

用户上下文

Why they needed it, what problem they're solving

他们的需求原因、要解决的问题

Complexity Estimate

复杂度评估

simple | medium | complex

Suggested Implementation

建议实现方案

How this could be built, what it might extend

如何构建、可扩展现有哪些功能

Metadata

元数据

Frequency: first_time | recurring
Related Features: existing_feature_name

undefined

频率: first_time | recurring
相关功能: existing_feature_name

undefined

ID Generation

ID生成规则

Format:

TYPE-YYYYMMDD-XXX

TYPE:
```
LRN
```
(learning),
```
ERR
```
(error),
```
FEAT
```
(feature)
YYYYMMDD: Current date
XXX: Sequential number or random 3 chars (e.g.,
```
001
```
,
```
A7B
```
)

Examples:

LRN-20250115-001

ERR-20250115-A3F

FEAT-20250115-002

格式:

TYPE-YYYYMMDD-XXX

TYPE:
```
LRN
```
（学习）、
```
ERR
```
（错误）、
```
FEAT
```
（功能）
YYYYMMDD: 当前日期
XXX: 序号或随机3字符（例如
```
001
```
、
```
A7B
```
）

示例:

LRN-20250115-001

ERR-20250115-A3F

FEAT-20250115-002

Resolving Entries

条目处理

When an issue is fixed, update the entry:

Change
```
**Status**: pending
```
→
```
**Status**: resolved
```
Add resolution block after Metadata:

markdown

undefined

当问题被修复后，更新条目：

将
```
**Status**: pending
```
改为
```
**Status**: resolved
```
在元数据后添加处理模块：

markdown

undefined

Resolution

处理详情

Resolved: 2025-01-16T09:00:00Z
Commit/PR: abc123 or #42
Notes: Brief description of what was done


Other status values:
- `in_progress` - Actively being worked on
- `wont_fix` - Decided not to address (add reason in Resolution notes)
- `promoted` - Elevated to CLAUDE.md, AGENTS.md, or .github/copilot-instructions.md

解决时间: 2025-01-16T09:00:00Z
提交/PR: abc123 或 #42
备注: 简要描述处理内容


其他状态值：
- `in_progress` - 正在处理中
- `wont_fix` - 决定不处理（在处理备注中添加原因）
- `promoted` - 已提升至CLAUDE.md、AGENTS.md或.github/copilot-instructions.md

Promoting to Project Memory

提升至项目记忆库

When a learning is broadly applicable (not a one-off fix), promote it to permanent project memory.

当学习内容具有广泛适用性（非一次性修复）时，将其提升至永久项目记忆库。

When to Promote

提升时机

Learning applies across multiple files/features
Knowledge any contributor (human or AI) should know
Prevents recurring mistakes
Documents project-specific conventions

学习内容适用于多个文件/功能
所有贡献者（人类或AI）都应了解的知识
可防止重复犯错
记录项目特定的约定

Promotion Targets

提升目标

Target	What Belongs There
`CLAUDE.md`	Project facts, conventions, gotchas for all Claude interactions
`AGENTS.md`	Agent-specific workflows, tool usage patterns, automation rules
`.github/copilot-instructions.md`	Project context and conventions for GitHub Copilot
`SOUL.md`	Behavioral guidelines, communication style, principles (OpenClaw workspace)
`TOOLS.md`	Tool capabilities, usage patterns, integration gotchas (OpenClaw workspace)

目标文件	适用内容
`CLAUDE.md`	项目事实、约定、所有Claude交互的注意事项
`AGENTS.md`	Agent特定工作流、工具使用模式、自动化规则
`.github/copilot-instructions.md`	GitHub Copilot的项目上下文和约定
`SOUL.md`	行为准则、沟通风格、原则（OpenClaw工作区）
`TOOLS.md`	工具能力、使用模式、集成注意事项（OpenClaw工作区）

How to Promote

提升步骤

Distill the learning into a concise rule or fact
Add to appropriate section in target file (create file if needed)

Update original entry:

Change
```
**Status**: pending
```
→
```
**Status**: promoted
```

Add

**Promoted**: CLAUDE.md

AGENTS.md

, or

.github/copilot-instructions.md

提炼：将学习内容浓缩为简洁的规则或事实
添加：添加至目标文件的对应章节（如文件不存在则创建）

更新原始条目：

将
```
**Status**: pending
```
改为
```
**Status**: promoted
```

添加

**Promoted**: CLAUDE.md

、

AGENTS.md

或

.github/copilot-instructions.md

Promotion Examples

提升示例

Learning (verbose):

Project uses pnpm workspaces. Attempted
npm install
but failed. Lock file is
pnpm-lock.yaml
. Must use
pnpm install
.

In CLAUDE.md (concise):

markdown

undefined

学习内容（详细版）：

项目使用pnpm工作区。尝试执行
npm install
失败。锁定文件为
pnpm-lock.yaml
。必须使用
pnpm install
。

在CLAUDE.md中（简洁版）：

markdown

undefined

Build & Dependencies

构建与依赖

Package manager: pnpm (not npm) - use
```
pnpm install
```


**Learning** (verbose):
> When modifying API endpoints, must regenerate TypeScript client.
> Forgetting this causes type mismatches at runtime.

**In AGENTS.md** (actionable):
```markdown

包管理器：pnpm（非npm）- 使用
```
pnpm install
```


**学习内容**（详细版）：
> 修改API端点后，必须重新生成TypeScript客户端。
> 忘记此步骤会导致运行时类型不匹配。

**在AGENTS.md中**（可操作版）：
```markdown

After API Changes

API修改后步骤

Regenerate client:
```
pnpm run generate:api
```
Check for type errors:
```
pnpm tsc --noEmit
```

undefined

重新生成客户端：
```
pnpm run generate:api
```
检查类型错误：
```
pnpm tsc --noEmit
```

undefined

Recurring Pattern Detection

重复模式检测

If logging something similar to an existing entry:

Search first:
```
grep -r "keyword" .learnings/
```
Link entries: Add
```
**See Also**: ERR-20250110-001
```
in Metadata
Bump priority if issue keeps recurring
Consider systemic fix: Recurring issues often indicate:
- Missing documentation (→ promote to CLAUDE.md or .github/copilot-instructions.md)
- Missing automation (→ add to AGENTS.md)
- Architectural problem (→ create tech debt ticket)

如果要记录的内容与现有条目相似：

先搜索：
```
grep -r "keyword" .learnings/
```
链接条目：在元数据中添加
```
**See Also**: ERR-20250110-001
```
提升优先级：如果问题重复出现
考虑系统性修复：重复问题通常表明：
- 缺少文档（→ 提升至CLAUDE.md或.github/copilot-instructions.md）
- 缺少自动化（→ 添加至AGENTS.md）
- 架构问题（→ 创建技术债务工单）

Periodic Review

定期回顾

Review

.learnings/

at natural breakpoints:

在合适的节点回顾

.learnings/

内容：

When to Review

回顾时机

Before starting a new major task
After completing a feature
When working in an area with past learnings
Weekly during active development

开始新的重大任务前
完成一个功能后
在有过往学习内容的领域工作时
活跃开发期间每周回顾

Quick Status Check

快速状态检查

bash

undefined

bash

undefined

Count pending items

统计待处理项数量

grep -h "Status**: pending" .learnings/*.md | wc -l

List pending high-priority items

列出待处理的高优先级项

grep -B5 "Priority**: high" .learnings/*.md | grep "^## ["

Find learnings for a specific area

查找特定领域的学习内容

grep -l "Area**: backend" .learnings/*.md

undefined

grep -l "Area**: backend" .learnings/*.md

undefined

Review Actions

回顾操作

Resolve fixed items
Promote applicable learnings
Link related entries
Escalate recurring issues

标记已修复的项为已解决
提升适用的学习内容
链接相关条目
升级重复出现的问题

Detection Triggers

检测触发条件

Automatically log when you notice:

Corrections (→ learning with

correction

category):

"No, that's not right..."
"Actually, it should be..."
"You're wrong about..."
"That's outdated..."

Feature Requests (→ feature request):

"Can you also..."
"I wish you could..."
"Is there a way to..."
"Why can't you..."

Knowledge Gaps (→ learning with

knowledge_gap

category):

User provides information you didn't know
Documentation you referenced is outdated
API behavior differs from your understanding

Errors (→ error entry):

Command returns non-zero exit code
Exception or stack trace
Unexpected output or behavior
Timeout or connection failure

当你注意到以下情况时自动记录：

修正内容（→ 分类为

correction

的学习条目）：

“不对，不是这样的……”
“实际上，应该是……”
“你错了……”
“这已经过时了……”

功能请求（→ 功能请求条目）：

“你还能……吗？”
“我希望你能……”
“有没有办法……？”
“为什么你不能……？”

知识缺口（→ 分类为

knowledge_gap

的学习条目）：

用户提供了你不知道的信息
你参考的文档已过时
API行为与你的理解不符

错误（→ 错误条目）：

命令返回非零退出码
异常或堆栈跟踪
意外输出或行为
超时或连接失败

Priority Guidelines

优先级指南

Priority	When to Use
`critical`	Blocks core functionality, data loss risk, security issue
`high`	Significant impact, affects common workflows, recurring issue
`medium`	Moderate impact, workaround exists
`low`	Minor inconvenience, edge case, nice-to-have

优先级	适用场景
`critical`	阻碍核心功能、存在数据丢失风险、安全问题
`high`	影响重大、影响常见工作流、重复出现的问题
`medium`	影响中等、存在替代方案
`low`	轻微不便、边缘情况、锦上添花的功能

Area Tags

领域标签

Use to filter learnings by codebase region:

Area	Scope
`frontend`	UI, components, client-side code
`backend`	API, services, server-side code
`infra`	CI/CD, deployment, Docker, cloud
`tests`	Test files, testing utilities, coverage
`docs`	Documentation, comments, READMEs
`config`	Configuration files, environment, settings

用于按代码库区域过滤学习内容：

领域	范围
`frontend`	UI、组件、客户端代码
`backend`	API、服务、服务器端代码
`infra`	CI/CD、部署、Docker、云服务
`tests`	测试文件、测试工具、覆盖率
`docs`	文档、注释、README
`config`	配置文件、环境、设置

Best Practices

最佳实践

Log immediately - context is freshest right after the issue
Be specific - future agents need to understand quickly
Include reproduction steps - especially for errors
Link related files - makes fixes easier
Suggest concrete fixes - not just "investigate"
Use consistent categories - enables filtering
Promote aggressively - if in doubt, add to CLAUDE.md or .github/copilot-instructions.md
Review regularly - stale learnings lose value

立即记录 - 问题刚发生时上下文最清晰
具体明确 - 未来的Agent需要快速理解
包含复现步骤 - 尤其是错误条目
链接相关文件 - 便于修复
建议具体的修复方案 - 不只是“调查”
使用一致的分类 - 便于过滤
积极提升内容 - 如有疑问，添加至CLAUDE.md或.github/copilot-instructions.md
定期回顾 - 过时的学习内容会失去价值

Gitignore Options

Gitignore选项

Keep learnings local (per-developer):

gitignore

.learnings/

Track learnings in repo (team-wide): Don't add to .gitignore - learnings become shared knowledge.

Hybrid (track templates, ignore entries):

gitignore

.learnings/*.md
!.learnings/.gitkeep

本地保留学习内容（每位开发者）：

gitignore

.learnings/

在仓库中跟踪学习内容（团队共享）：不要添加到.gitignore - 学习内容成为共享知识。

混合模式（跟踪模板，忽略条目）：

gitignore

.learnings/*.md
!.learnings/.gitkeep

Hook Integration

钩子集成

Enable automatic reminders through agent hooks. This is opt-in - you must explicitly configure hooks.

通过Agent钩子启用自动提醒。此功能为可选 - 你必须显式配置钩子。

Quick Setup (Claude Code / Codex)

快速配置（Claude Code / Codex）

Create

.claude/settings.json

in your project:

json

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/activator.sh"
      }]
    }]
  }
}

This injects a learning evaluation reminder after each prompt (~50-100 tokens overhead).

在你的项目中创建

.claude/settings.json

：

json

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/activator.sh"
      }]
    }]
  }
}

这会在每个提示后注入学习评估提醒（约50-100令牌开销）。

Full Setup (With Error Detection)

完整配置（含错误检测）

json

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/activator.sh"
      }]
    }],
    "PostToolUse": [{
      "matcher": "Bash",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/error-detector.sh"
      }]
    }]
  }
}

json

{
  "hooks": {
    "UserPromptSubmit": [{
      "matcher": "",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/activator.sh"
      }]
    }],
    "PostToolUse": [{
      "matcher": "Bash",
      "hooks": [{
        "type": "command",
        "command": "./skills/self-improvement/scripts/error-detector.sh"
      }]
    }]
  }
}

Available Hook Scripts

可用钩子脚本

Script	Hook Type	Purpose
`scripts/activator.sh`	UserPromptSubmit	Reminds to evaluate learnings after tasks
`scripts/error-detector.sh`	PostToolUse (Bash)	Triggers on command errors

See

references/hooks-setup.md

for detailed configuration and troubleshooting.

脚本	钩子类型	用途
`scripts/activator.sh`	UserPromptSubmit	任务完成后提醒评估学习内容
`scripts/error-detector.sh`	PostToolUse (Bash)	命令错误时触发

详细配置与故障排除请参见

references/hooks-setup.md

。

Automatic Skill Extraction

自动Skill提取

When a learning is valuable enough to become a reusable skill, extract it using the provided helper.

当学习内容足够有价值可成为可复用Skill时，使用提供的工具提取。

Skill Extraction Criteria

Skill提取标准

A learning qualifies for skill extraction when ANY of these apply:

Criterion	Description
Recurring	Has `See Also` links to 2+ similar issues
Verified	Status is `resolved` with working fix
Non-obvious	Required actual debugging/investigation to discover
Broadly applicable	Not project-specific; useful across codebases
User-flagged	User says "save this as a skill" or similar

当满足以下任一条件时，学习内容可提取为Skill：

标准	描述
重复出现	有 `See Also` 链接至2个以上相似问题
已验证	状态为 `resolved` 且修复方案有效
非显而易见	需要实际调试/调查才能发现
广泛适用	非项目特定；适用于多个代码库
用户标记	用户说“将此保存为Skill”或类似表述

Extraction Workflow

提取流程

Identify candidate: Learning meets extraction criteria

Run helper (or create manually):

bash

./skills/self-improvement/scripts/extract-skill.sh skill-name --dry-run
./skills/self-improvement/scripts/extract-skill.sh skill-name

Customize SKILL.md: Fill in template with learning content
Update learning: Set status to
```
promoted_to_skill
```
, add
```
Skill-Path
```
Verify: Read skill in fresh session to ensure it's self-contained

识别候选内容：学习内容满足提取标准

运行工具（或手动创建）：

bash

./skills/self-improvement/scripts/extract-skill.sh skill-name --dry-run
./skills/self-improvement/scripts/extract-skill.sh skill-name

自定义SKILL.md：使用模板填充学习内容
更新学习条目：设置状态为
```
promoted_to_skill
```
，添加
```
Skill-Path
```
验证：在新会话中读取Skill，确保其独立可用

Manual Extraction

手动提取

If you prefer manual creation:

Create
```
skills/<skill-name>/SKILL.md
```
Use template from
```
assets/SKILL-TEMPLATE.md
```
Follow Agent Skills spec:
- YAML frontmatter with
```
name
```
  and
```
description
```
- Name must match folder name
- No README.md inside skill folder

如果你偏好手动创建：

创建
```
skills/<skill-name>/SKILL.md
```
使用
```
assets/SKILL-TEMPLATE.md
```
中的模板
遵循Agent Skills规范：
- 包含
```
name
```
  和
```
description
```
  的YAML前置元数据
- 名称必须与文件夹名称匹配
- Skill文件夹内无README.md

Extraction Detection Triggers

提取检测触发条件

Watch for these signals that a learning should become a skill:

In conversation:

"Save this as a skill"
"I keep running into this"
"This would be useful for other projects"
"Remember this pattern"

In learning entries:

Multiple
```
See Also
```
links (recurring issue)
High priority + resolved status
Category:
```
best_practice
```
with broad applicability
User feedback praising the solution

注意以下信号，表明学习内容应转化为Skill：

对话中：

“将此保存为Skill”
“我总是遇到这个问题”
“这对其他项目也有用”
“记住这个模式”

学习条目中：

多个
```
See Also
```
链接（重复问题）
高优先级+已解决状态
分类为
```
best_practice
```
且具有广泛适用性
用户反馈称赞解决方案

Skill Quality Gates

Skill质量把关

Multi-Agent Support

多Agent支持

This skill works across different AI coding agents with agent-specific activation.

此Skill可在不同的AI编码Agent中使用，每个Agent有特定的激活方式。

Claude Code

Activation: Hooks (UserPromptSubmit, PostToolUse) Setup:

.claude/settings.json

with hook configuration Detection: Automatic via hook scripts

激活方式：钩子（UserPromptSubmit、PostToolUse）配置：

.claude/settings.json

中的钩子配置 检测方式：通过钩子脚本自动检测

Codex CLI

Activation: Hooks (same pattern as Claude Code) Setup:

.codex/settings.json

with hook configuration Detection: Automatic via hook scripts

激活方式：钩子（与Claude Code模式相同）配置：

.codex/settings.json

中的钩子配置 检测方式：通过钩子脚本自动检测

GitHub Copilot

Activation: Manual (no hook support) Setup: Add to

.github/copilot-instructions.md

markdown

undefined

激活方式：手动（无钩子支持）配置：添加至

.github/copilot-instructions.md

：

markdown

undefined

Self-Improvement

自我改进

After solving non-obvious issues, consider logging to

.learnings/

Use format from self-improvement skill
Link related entries with See Also
Promote high-value learnings to skills

Ask in chat: "Should I log this as a learning?"


**Detection**: Manual review at session end

解决非显而易见的问题后，考虑将内容记录至

.learnings/

：

使用自我改进Skill的格式
使用See Also链接相关条目
将高价值学习内容提升为Skill

在聊天中询问：“我应该将此记录为学习内容吗？”


**检测方式**：会话结束时手动回顾

OpenClaw

Activation: Workspace injection + inter-agent messaging Setup: See "OpenClaw Setup" section above Detection: Via session tools and workspace files

激活方式：工作区注入+Agent间通信配置：参见上方“OpenClaw配置”部分 检测方式：通过会话工具和工作区文件

Agent-Agnostic Guidance

Agent通用指南

Regardless of agent, apply self-improvement when you:

Discover something non-obvious - solution wasn't immediate
Correct yourself - initial approach was wrong
Learn project conventions - discovered undocumented patterns
Hit unexpected errors - especially if diagnosis was difficult
Find better approaches - improved on your original solution

无论使用哪种Agent，在以下场景应用自我改进：

发现非显而易见的内容 - 解决方案并非立即可得
自我纠正 - 初始方法错误
学习项目约定 - 发现未文档化的模式
遇到意外错误 - 尤其是诊断困难的情况
找到更优方法 - 改进原始解决方案

Copilot Chat Integration

Copilot Chat集成

For Copilot users, add this to your prompts when relevant:

After completing this task, evaluate if any learnings should be logged to
.learnings/
using the self-improvement skill format.

Or use quick prompts:

"Log this to learnings"
"Create a skill from this solution"
"Check .learnings/ for related issues"

对于Copilot用户，在相关提示中添加以下内容：

完成此任务后，评估是否有任何学习内容应使用自我改进Skill的格式记录至
.learnings/
。

或使用快速提示：

“将此记录至学习内容”
“从此解决方案创建Skill”
“检查.learnings/中的相关问题”