OpenClaw

metaskill

Teaches AI agents how to learn better by enforcing deep correction, transfer learning, and proactive pattern recognition. Use when an error occurs and needs deep analysis (not surface patch), before starting a complex task to check past analogies, or after success to capture winning patterns. NOT for routine file reads or simple one-off commands.

fahrulalwan0 来源安装数☆ 0 仓库 Stars

查看来源 ↗安装与使用 SKILL.md ↗

安装与使用

请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求，再添加到工作流。

SkillHub CLI

skillhub install fahrulalwan/metaskill

打开来源仓库，查看说明和兼容环境后，按来源提供的命令或步骤安装。

将来源链接提供给你的 Agent，让它先读取说明、核对依赖，并向你说明将执行的安装步骤。

查看来源 ↗使用指南

技能说明

Metaskill

3 Core Components

Deep Self-Correction (deep-correct.sh) — 3-level breakdown on errors:
- Surface: What specifically failed
- Principle: The underlying rule/constraint violated
- Habit: Concrete behavioral change to prevent recurrence
Transfer Learning (transfer-check.sh) — Before a task, search past learnings for analogous patterns. Maps domains (e.g., "auth" → "security") to prevent siloed learning.
Proactive Pattern Recognition (success-capture.sh) — Log what worked and why, building a repository of successful patterns.

Usage

# When an error occurs
bash skills/metaskill/scripts/deep-correct.sh "description of the error"

# Before starting a complex task
bash skills/metaskill/scripts/transfer-check.sh "description of the new task"

# After successful execution
bash skills/metaskill/scripts/success-capture.sh "what worked" "why it worked"

# Monthly health eval
bash skills/metaskill/scripts/eval.sh --save

Configuration (LLM Provider)

Metaskill uses two provider tiers — fast (extraction) and deep (transfer/eval). Edit config.yaml to match your setup:

# config.yaml
providers:
  fast: anthropic   # change to: openai | ollama | gemini
  deep: anthropic

Provider	Env Var	Notes
`anthropic`	`ANTHROPIC_API_KEY`	Default
`openai`	`OPENAI_API_KEY`
`ollama`	(none needed)	Local, free
`gemini`	`GOOGLE_API_KEY`

Ollama example (fully local, no API key):

providers:
  fast: ollama
  deep: ollama
models:
  ollama:
    fast: llama3.2
    deep: llama3.1:70b

If no provider is available, metaskill falls back to manual/heuristic mode (still works, but less precise extraction).

Integration with Self-Improving-Agent

Writes to skills/self-improving-agent/.learnings/ if present, otherwise falls back to its own .learnings/ directory. No extra setup needed.

AGENTS.md Wiring (Mandatory)

Add to pre-t...

来源信息

Source: github
Repository: openclaw/skills
Path: skills/fahrulalwan/metaskill
Version: —
Commit: 38d8bf440ce2960d6dab4408345fccc750d91d12
分类来源: rules
状态: active
可见性: public

这些信息来自收录时的来源分析，可能滞后；使用前请核对最新来源文件。

来源内容

SKILL.md

# Metaskill

## 3 Core Components

1. **Deep Self-Correction (`deep-correct.sh`)** — 3-level breakdown on errors:
   - **Surface**: What specifically failed
   - **Principle**: The underlying rule/constraint violated
   - **Habit**: Concrete behavioral change to prevent recurrence

2. **Transfer Learning (`transfer-check.sh`)** — Before a task, search past learnings for analogous patterns. Maps domains (e.g., "auth" → "security") to prevent siloed learning.

3. **Proactive Pattern Recognition (`success-capture.sh`)** — Log what worked and why, building a repository of successful patterns.

## Usage

```bash
# When an error occurs
bash skills/metaskill/scripts/deep-correct.sh "description of the error"

# Before starting a complex task
bash skills/metaskill/scripts/transfer-check.sh "description of the new task"

# After successful execution
bash skills/metaskill/scripts/success-capture.sh "what worked" "why it worked"

# Monthly health eval
bash skills/metaskill/scripts/eval.sh --save
```

## Configuration (LLM Provider)

Metaskill uses two provider tiers — **fast** (extraction) and **deep** (transfer/eval). Edit `config.yaml` to match your setup:

```yaml
# config.yaml
providers:
  fast: anthropic   # change to: openai | ollama | gemini
  deep: anthropic
```

| Provider | Env Var | Notes |
|---|---|---|
| `anthropic` | `ANTHROPIC_API_KEY` | Default |
| `openai` | `OPENAI_API_KEY` | |
| `ollama` | *(none needed)* | Local, free |
| `gemini` | `GOOGLE_API_KEY` | |

**Ollama example** (fully local, no API key):
```yaml
providers:
  fast: ollama
  deep: ollama
models:
  ollama:
    fast: llama3.2
    deep: llama3.1:70b
```

If no provider is available, metaskill falls back to manual/heuristic mode (still works, but less precise extraction).

## Integration with Self-Improving-Agent

Writes to `skills/self-improving-agent/.learnings/` if present, otherwise falls back to its own `.learnings/` directory. No extra setup needed.

## AGENTS.md Wiring (Mandatory)

Add to pre-t...

搜索