mcpField
S
Self-Evolving Skill

Self-Evolving Skill

github.com/jackjin1997Agent workflows13315 installsUpdated 19w ago
Repository

About Self-Evolving Skill

Self-Evolving Skill is a self-learning system for evolving Skills through predictive coding and value-driven updates. It is intended for building and managing Skills that adapt from execution experience, combining gap analysis, reflection triggers, replay, and persistence.

Key features

  1. 01Residual-pyramid decomposition and gap measurement
  2. 02Adaptive reflection triggers based on residual energy
  3. 03Replay cache for learned patterns
  4. 04Value gate for filtering mutations
  5. 05Persistent experience storage and loading

Use cases

  1. 01Adjust Skills based on execution results
  2. 02Analyze embeddings and assess novelty
  3. 03Create, execute, and list Skills
  4. 04Save or load Skill experience

MCP support

It provides an MCP server exposing tools to create, execute, analyze, list, and inspect Skills, as well as save and load them.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/jackjin1997/clawforge --skill self-evolving-skill

After installing, type/self-evolving-skillin Claude Code to confirm it loaded.

Details

From the project README

Self-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Skillname, 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.jsonMCP服务器配置
~/.openclaw/workspace/self-evolving-skill/storage数据存储

相关文档

Comments

Markdown supported