Meta Skill Optimizer
Self-improving AI capability that enables continuous skill enhancement.
Features
1. Feedback Learning
- - Success Analysis: Learn from successful executions
- Failure Analysis: Understand and prevent failures
- Pattern Recognition: Identify recurring patterns
- Preference Learning: Adapt to user preferences
2. Prompt Optimization
- - Auto-Tuning: Optimize prompts based on outcomes
- Chain-of-Thought: Improve reasoning chains
- Example Selection: Dynamic few-shot example selection
- Style Adaptation: Match user communication style
3. Tool Usage Optimization
- - Tool Selection: Choose best tools for tasks
- Parameter Tuning: Optimize tool parameters
- Workflow Patterns: Discover effective workflows
- Error Recovery: Learn from tool errors
4. Self-Diagnosis
- - Capability Assessment: Know what it can/can't do
- Knowledge Gaps: Identify missing knowledge
- Confidence Calibration: Accurate confidence levels
- Limitation Awareness: Know when to ask for help
5. Continuous Evolution
- - Version Tracking: Track skill improvements
- A/B Testing: Compare approach effectiveness
- Best Practices: Extract and codify learnings
- Knowledge Base: Build searchable knowledge
Installation
CODEBLOCK0
Usage
Initialize Optimizer
CODEBLOCK1
Record Execution Result
CODEBLOCK2
Get Optimized Approach
CODEBLOCK3
Optimize Prompt
CODEBLOCK4
API Reference
Feedback Learning
| Method | Description |
|---|
| INLINECODE0 | Record successful execution |
| INLINECODE1 |
Record failed execution |
|
get_insights() | Get learned insights |
Prompt Optimization
| Method | Description |
|---|
| INLINECODE3 | Optimize prompt based on feedback |
| INLINECODE4 |
Generate few-shot examples |
|
adapt_style(...) | Adapt to user style |
Tool Optimization
| Method | Description |
|---|
| INLINECODE6 | Suggest best tools |
| INLINECODE7 |
Optimize tool parameters |
|
discover_workflow(...) | Discover effective workflows |
Self-Diagnosis
| Method | Description |
|---|
| INLINECODE9 | Assess capability for task |
| INLINECODE10 |
Identify knowledge gaps |
|
calibrate_confidence() | Calibrate confidence levels |
Evolution
| Method | Description |
|---|
| INLINECODE12 | Track improvement over time |
| INLINECODE13 |
Export learned knowledge |
|
merge_experiences() | Merge from other optimizers |
How It Works
1. Feedback Loop
CODEBLOCK5
2. Pattern Discovery
CODEBLOCK6
3. Continuous Learning
CODEBLOCK7
Use Cases
- - Prompt Engineering: Continuously improve prompts
- Tool Selection: Better tool recommendations
- Error Prevention: Learn from past mistakes
- User Adaptation: Match user preferences
- Capability Growth: Expand what AI can do
Knowledge Base
The optimizer builds a knowledge base:
CODEBLOCK8
Integration
With OpenClaw
CODEBLOCK9
With Skills
CODEBLOCK10
Best Practices
- 1. Record Everything: More data = better learning
- Categorize Failures: Understand failure types
- Update Regularly: Keep knowledge current
- Merge Insights: Combine learnings from multiple sources
Future Capabilities
- - Cross-skill learning
- Automatic skill creation
- Self-debugging
- Automated testing
元技能优化器
具备自我改进能力的AI，实现技能的持续增强。
特性
1. 反馈学习
- - 成功分析：从成功执行中学习
- 失败分析：理解并预防失败
- 模式识别：识别重复出现的模式
- 偏好学习：适应用户偏好
2. 提示词优化
- - 自动调优：基于结果优化提示词
- 思维链：改进推理链条
- 示例选择：动态选择少量示例
- 风格适配：匹配用户沟通风格
3. 工具使用优化
- - 工具选择：为任务选择最佳工具
- 参数调优：优化工具参数
- 工作流模式：发现高效工作流
- 错误恢复：从工具错误中学习
4. 自我诊断
- - 能力评估：了解自身能/不能做什么
- 知识缺口：识别缺失的知识
- 置信度校准：准确的置信度水平
- 局限意识：知道何时寻求帮助
5. 持续进化
- - 版本追踪：跟踪技能改进
- A/B测试：比较方法有效性
- 最佳实践：提取并编码学习成果
- 知识库：构建可搜索的知识体系
安装
bash
pip install numpy scipy json
使用
初始化优化器
python
from meta_optimizer import SkillOptimizer
optimizer = SkillOptimizer(
skillname=dataanalysis,
learning_rate=0.1
)
记录执行结果
python
记录成功执行
optimizer.record_success(
task=analyze sales data,
approach=used pandas groupby,
context={data_size: 10MB, complexity: high},
outcome={success: True, quality: high}
)
记录失败
optimizer.record_failure(
task=predict stock price,
approach=used linear regression,
error=insufficient features,
lesson=need more technical indicators
)
获取优化后的方法
python
获取任务的最佳方法
best
approach = optimizer.getbest_approach(
task
type=dataanalysis,
context={data_size: 1GB}
)
print(best_approach)
{method: chunked_processing, tools: [pandas, dask]}
优化提示词
python
基于结果优化提示词
optimized
prompt = optimizer.optimizeprompt(
original_prompt=Analyze this data,
outcome=too vague,
feedback=be more specific about analysis type
)
print(optimized_prompt)
Analyze this time-series data using trend detection and seasonality analysis
API参考
反馈学习
| 方法 | 描述 |
|---|
| recordsuccess(...) | 记录成功执行 |
| recordfailure(...) |
记录失败执行 |
| get_insights() | 获取学习到的见解 |
提示词优化
| 方法 | 描述 |
|---|
| optimizeprompt(...) | 基于反馈优化提示词 |
| generateexamples(...) |
生成少量示例 |
| adapt_style(...) | 适应用户风格 |
工具优化
| 方法 | 描述 |
|---|
| suggesttools(...) | 推荐最佳工具 |
| optimizeparams(...) |
优化工具参数 |
| discover_workflow(...) | 发现高效工作流 |
自我诊断
| 方法 | 描述 |
|---|
| assesscapability(...) | 评估任务能力 |
| identifygaps() |
识别知识缺口 |
| calibrate_confidence() | 校准置信度水平 |
进化
| 方法 | 描述 |
|---|
| trackimprovement() | 追踪随时间改进 |
| exportknowledge() |
导出学到的知识 |
| merge_experiences() | 从其他优化器合并经验 |
工作原理
1. 反馈循环
任务 → 执行 → 结果 → 反馈 → 学习 → 改进
2. 模式发现
多次执行 → 模式挖掘 → 最佳实践 → 编码
3. 持续学习
新任务 → 相似历史任务 → 学到的经验 → 优化方法
使用场景
- - 提示词工程：持续改进提示词
- 工具选择：更好的工具推荐
- 错误预防：从过去错误中学习
- 用户适配：匹配用户偏好
- 能力增长：扩展AI能力范围
知识库
优化器构建知识库：
json
{
patterns: {
data_analysis: {
small_data: pandas sufficient,
large_data: use dask or chunking,
time_series: check stationarity first
}
},
prompts: {
effective: [specific, contextual, actionable],
ineffective: [vague, ambiguous, overly broad]
},
tools: {
coding: [cursor, claude-code],
research: [tavily, browser]
}
}
集成
与OpenClaw集成
python
自动记录所有执行
@hookimpl
def after_execution(result, context):
optimizer.record_execution(context, result)
与技能集成
python
优化技能行为
skill = MySkill()
optimized
skill = optimizer.optimizeskill(skill)
最佳实践
- 1. 记录一切：更多数据 = 更好学习
- 分类失败：理解失败类型
- 定期更新：保持知识时效性
- 合并见解：结合多个来源的学习成果
未来能力
