Self Evolving Skill
Meta-cognitive self-learning system - Automated skill evolution based on predictive coding and value-driven mechanisms.
bltdz Orion
@whtoo
What This Skill Does
Meta-cognitive self-learning system that automatically evolves skills based on predictive coding and value-driven mechanisms. It uses residual pyramid decomposition to quantify knowledge gaps, triggers learning when novelty is detected, and only accepts mutations that increase long-term value.
Replaces manual skill tuning and static rule-based systems by enabling autonomous, adaptive skill evolution driven by predictive error and value gates.
When to Use It
- Automatically refine an AI agent's skills based on performance feedback
- Detect and fill knowledge gaps in an agent's understanding of a domain
- Evolve a skill's behavior to maximize cumulative reward over time
- Create new sub-skills or predicates when existing ones cover less than 40% of cases
- Persist learned patterns and reload them across agent sessions
- Analyze embedding novelty to decide whether to trigger a learning cycle
Install
$ openclaw skills install @whtoo/self-evolving-skillSelf-Evolving Skill
元认知自学习系统 - 基于预测编码和价值驱动的Skill自动演化。
功能
- ResidualPyramid金字塔分解,量化认知缺口 -: 残差 自适应反思触发: 基于残差能量自动判断何时需要学习
- 经验回放: 缓存已学模式,降低重复触发
- 价值门控: 只有提升长期价值才接受变异
- 持久化: 经验自动保存/加载
安装
# 技能已安装到 ~/.openclaw/skills/self-evolving-skill
# 或使用ClawHub
clawhub install self-evolving-skill
架构
self-evolving-skill/
├── core/ # Python核心
│ ├── residual_pyramid.py # 残差金字塔(SVD分解)
│ ├── reflection_trigger.py # 自适应触发器
│ ├── experience_replay.py # 经验回放缓存
│ ├── skill_engine.py # 核心引擎+ValueGate
│ ├── storage.py # 持久化
│ └── mcp_server.py # MCP服务器
├── src/ # TypeScript SDK
│ ├── index.ts # 主入口
│ ├── cli.ts # CLI
│ └── mcp-tools.ts # 工具定义
├── skills/ # OpenClaw Skill
│ └── self-evolving-skill/ # 技能封装
├── MCP_CONFIG.md # MCP配置
└── README.md # 文档
MCP工具
| 工具 | 描述 | 参数 |
|---|---|---|
skill_create | 创建Skill | name, description |
skill_execute | 执行并学习 | skill_id, context, success, value |
skill_analyze | 分析嵌入 | embedding |
skill_list | 列出Skills | - |
skill_stats | 系统统计 | - |
skill_save | 持久化保存 | skill_id |
skill_load | 加载 | skill_id |
使用方式
CLI
# 列出所有Skill
openclaw skill self-evolving-skill list
# 创建Skill
openclaw skill self-evolving-skill create --name "MySkill"
# 执行
openclaw skill self-evolving-skill execute <id> --success
# 分析
openclaw skill self-evolving-skill analyze --embedding '[0.1,0.2,...]'
# 统计
openclaw skill self-evolving-skill stats
MCP服务器
# 启动MCP服务器
cd ~/.openclaw/skills/self-evolving-skill
./run_mcp.sh
# 或使用适配器
python3 mcporter_adapter.py skill_list '{}'
编程
import { SelfEvolvingSkillEngine } from 'self-evolving-skill';
const engine = new SelfEvolvingSkillEngine();
await engine.init();
const { skillId } = await engine.createSkill({ name: 'Analyzer' });
const stats = await engine.stats();
核心算法
1. 残差金字塔分解
pyramid = ResidualPyramid(max_layers=5, use_pca=True)
decomposition = pyramid.decompose(embedding)
# 输出:
# - residual_ratio: 残差能量比率
# - suggested_abstraction: POLICY / SUB_SKILL / PREDICATE
# - novelty_score: 综合新颖性
2. 三层跃迁规则
| 覆盖率 | 抽象层级 | 操作 |
|---|---|---|
| >80% | POLICY | 调整策略权重 |
| 40-80% | SUB_SKILL | 生成子Skill |
| <40% | PREDICATE | 归纳新谓词 |
3. 自适应阈值
trigger = ReflectionTrigger(
min_energy_ratio=0.10, # 初始阈值
value_gain_threshold=0.20, # 触发阈值
target_trigger_rate=0.15 # 目标15%触发率
)
文件位置
| 路径 | 说明 |
|---|---|
~/.openclaw/skills/self-evolving-skill | 技能根目录 |
~/.openclaw/mcp_servers/self-evolving-skill.json | MCP服务器配置 |
~/.openclaw/workspace/self-evolving-skill/storage | 数据存储 |
相关文档
- README.md - 完整文档
- MCP_CONFIG.md - MCP配置说明
- MEMORY.md - 研究笔记
Top skills in this category
Skill Vetter
@spclaudehomeSecurity-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
Self-Improving + Proactive Agent
@ivangdavilaSelf-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
Humanizer
@biostartechnologyRemove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.
Elite Longterm Memory
@nextfrontierbuildsUltimate AI agent memory system for Cursor, Claude, ChatGPT & Copilot. WAL protocol + vector search + git-notes + cloud backup. Never lose context again. Vibe-coding ready.
Clawdhub
@steipeteUse the ClawdHub CLI to search, install, update, and publish agent skills from clawdhub.com. Use when you need to fetch new skills on the fly, sync installed skills to latest or a specific version, or publish new/updated skill folders with the npm-installed clawdhub CLI.