Skill Learner

Automatically analyze completed work, extract reusable patterns, and create/enhance skills for continual organizational learning.

自动分析已完成的工作，提取可复用模式，并创建/增强技能以实现组织的持续学习。

Quick Start

快速开始

bash

undefined

bash

undefined

Manual trigger after commit

/skill-learner

Auto-triggers after:

- git commit (via post-commit hook)

- Task completion

- Feature implementation

Review learning

cat .claude/skill-learning-log.md

undefined

cat .claude/skill-learning-log.md

undefined

When to Use

适用场景

AUTO-EXECUTES (via hook):

After every git commit
After task completion
After feature implementation
After bug fixes with significant patterns
After refactoring work

MANUAL TRIGGER:

To analyze recent work history
To review learning opportunities
To create skills from existing patterns
After completing a project phase
During knowledge capture sessions

自动执行（通过钩子）：

每次git提交后
任务完成后
功能实现后
修复带有重要模式的bug后
重构工作后

手动触发：

分析近期工作历史
审视学习机会
从现有模式创建技能
完成项目阶段后
知识捕获会议期间

Prerequisites

前置条件

Git repository with commit history
Access to committed files and diffs
(Optional) Skills directory for new skill creation
(Optional) Internet for pattern research

带有提交历史的Git仓库
可访问已提交的文件和差异内容
（可选）用于创建新技能的技能目录
（可选）用于模式研究的网络连接

Overview

概述

The skill-learner performs post-commit analysis to identify reusable patterns, common workflows, and valuable techniques from completed work. It automatically creates new skills or enhances existing ones, ensuring continuous organizational learning.

Skill Learner执行提交后分析，从已完成的工作中识别可复用模式、通用工作流和有价值的技术。它会自动创建新技能或增强现有技能，确保组织的持续学习。

What It Analyzes

分析内容

Commit Content: Files changed, code added, patterns used
Commit Message: Task description, context, intent
Work Patterns: Repeated workflows, common operations
Problem Solutions: How bugs were fixed, how features were built
Tool Usage: Libraries, frameworks, techniques employed
Documentation: Comments, docs added, conventions followed

提交内容：修改的文件、添加的代码、使用的模式
提交信息：任务描述、上下文、意图
工作模式：重复的工作流、常见操作
问题解决方案：bug修复方式、功能构建方法
工具使用：采用的库、框架、技术
文档：添加的注释、文档、遵循的规范

Output

输出内容

Generates:

Pattern analysis report
Skill creation recommendations
Enhanced existing skills
Learning log with insights
Knowledge base entries

生成：

模式分析报告
技能创建建议
增强后的现有技能
包含见解的学习日志
知识库条目

Core Operations

核心操作

1. Commit Analysis

1. 提交分析

Analyzes Last Commit:

bash

undefined

分析最近一次提交：

bash

undefined

Extract commit metadata

COMMIT_HASH=$(git rev-parse HEAD) COMMIT_MSG=$(git log -1 --pretty=%B) COMMIT_AUTHOR=$(git log -1 --pretty=%an) COMMIT_DATE=$(git log -1 --pretty=%ai)

Get changed files

git diff-tree --no-commit-id --name-only -r HEAD

Get diff content

git diff HEAD^ HEAD


**Example Analysis:**
```markdown

git diff HEAD^ HEAD


**分析示例：**
```markdown

Commit Analysis

Commit: a1b2c3d4 Message: Add interactive NPV calculator with Plotly visualization Author: Developer Date: 2026-01-07 14:30:00

Files Changed (12)

src/modules/npv/calculator.py (new)
src/modules/npv/visualizer.py (new)
src/modules/npv/init.py (new)
tests/unit/test_npv_calculator.py (new)
config/input/npv_analysis.yaml (new)
scripts/run_npv_analysis.sh (new)
docs/npv_calculator.md (new)
... 5 more files

src/modules/npv/calculator.py (new)
src/modules/npv/visualizer.py (new)
src/modules/npv/init.py (new)
tests/unit/test_npv_calculator.py (new)
config/input/npv_analysis.yaml (new)
scripts/run_npv_analysis.sh (new)
docs/npv_calculator.md (new)
... 5 more files

Code Additions

Lines added: 847
Lines removed: 23
Net change: +824

Lines added: 847
Lines removed: 23
Net change: +824

Technologies Used

Plotly for interactive visualization
Pandas for data handling
NumPy for NPV calculations
pytest for testing

undefined

Plotly for interactive visualization
Pandas for data handling
NumPy for NPV calculations
pytest for testing

undefined

2. Pattern Extraction

2. 模式提取

Identifies Reusable Patterns:

Pattern Types:

markdown

1. **Workflow Patterns**
   - YAML config → Script execution → HTML report
   - Data load → Process → Validate → Visualize → Save

2. **Code Patterns**
   - Interactive plotting with Plotly
   - CSV data loading with relative paths
   - Modular calculator structure
   - TDD test-first approach

3. **Tool Patterns**
   - UV environment management
   - Bash script execution
   - YAML configuration files
   - Plotly visualization

4. **Problem-Solving Patterns**
   - NPV calculation with multiple discount rates
   - Handling missing data gracefully
   - Interactive parameter adjustment
   - Real-time visualization updates

Example Extraction:

markdown

undefined

识别可复用模式：

模式类型：

markdown

1. **Workflow Patterns**
   - YAML config → Script execution → HTML report
   - Data load → Process → Validate → Visualize → Save

2. **Code Patterns**
   - Interactive plotting with Plotly
   - CSV data loading with relative paths
   - Modular calculator structure
   - TDD test-first approach

3. **Tool Patterns**
   - UV environment management
   - Bash script execution
   - YAML configuration files
   - Plotly visualization

4. **Problem-Solving Patterns**
   - NPV calculation with multiple discount rates
   - Handling missing data gracefully
   - Interactive parameter adjustment
   - Real-time visualization updates

提取示例：

markdown

undefined

Pattern: Interactive Financial Calculator

Problem Solved

Need to calculate NPV with multiple scenarios and visualize results interactively.

Solution Pattern

YAML configuration with parameters
Python calculator module
Plotly visualization
Bash execution wrapper
HTML report generation

YAML configuration with parameters
Python calculator module
Plotly visualization
Bash execution wrapper
HTML report generation

Reusability Score: 95/100

Applicable to: IRR, ROI, payback period, sensitivity analysis
Generic enough: Yes
Well-documented: Yes
Tested: Yes (85% coverage)

undefined

Applicable to: IRR, ROI, payback period, sensitivity analysis
Generic enough: Yes
Well-documented: Yes
Tested: Yes (85% coverage)

undefined

3. Skill Identification

3. 技能识别

Determines Skill Creation Need:

Decision Matrix:

CREATE NEW SKILL if:
✓ Pattern used 3+ times across commits
✓ Workflow is complex (5+ steps)
✓ Domain-specific knowledge required
✓ Significant time savings (>30 min per use)
✓ No existing skill covers it

ENHANCE EXISTING SKILL if:
✓ Pattern similar to existing skill
✓ New technique for known problem
✓ Updated best practices
✓ New tool/library version
✓ Improved approach discovered

SKIP if:
✗ One-time solution
✗ Trivial pattern (<3 steps)
✗ Already well-covered
✗ Repository-specific only

Example Decision:

markdown

undefined

判断是否需要创建技能：

决策矩阵：

CREATE NEW SKILL if:
✓ Pattern used 3+ times across commits
✓ Workflow is complex (5+ steps)
✓ Domain-specific knowledge required
✓ Significant time savings (>30 min per use)
✓ No existing skill covers it

ENHANCE EXISTING SKILL if:
✓ Pattern similar to existing skill
✓ New technique for known problem
✓ Updated best practices
✓ New tool/library version
✓ Improved approach discovered

SKIP if:
✗ One-time solution
✗ Trivial pattern (<3 steps)
✗ Already well-covered
✗ Repository-specific only

决策示例：

markdown

undefined

Skill Decision: NPV Calculator

Analysis

Frequency: Used 5 times in last month
Complexity: 8-step workflow
Domain Knowledge: Financial engineering
Time Savings: ~2 hours per use
Existing Skills: None for financial calculators

Frequency: Used 5 times in last month
Complexity: 8-step workflow
Domain Knowledge: Financial engineering
Time Savings: ~2 hours per use
Existing Skills: None for financial calculators

Recommendation: CREATE NEW SKILL

Skill Name: financial-calculator-builder Category: development/finance Priority: High

Rationale:

High reusability across energy economic analysis
Complex enough to warrant skill documentation
Significant time savings
Establishes pattern for future financial tools

undefined

Skill Name: financial-calculator-builder Category: development/finance Priority: High

Rationale:

High reusability across energy economic analysis
Complex enough to warrant skill documentation
Significant time savings
Establishes pattern for future financial tools

undefined

4. Skill Creation

4. 技能创建

Automatically Creates Skills:

Creation Process:

bash

1. Generate skill name from pattern
2. Extract workflow steps
3. Create SKILL.md with template
4. Document code examples
5. Add usage instructions
6. Link related skills
7. Update skills README
8. Commit new skill

Example Skill Created:

markdown

---
name: financial-calculator-builder
description: Build interactive financial calculators (NPV, IRR, ROI) with Plotly visualization, YAML configuration, and HTML reporting.
version: 1.0.0
category: development/finance
created_from: commit a1b2c3d4
pattern_source: npv_calculator implementation
---

自动创建技能：

创建流程：

bash

1. Generate skill name from pattern
2. Extract workflow steps
3. Create SKILL.md with template
4. Document code examples
5. Add usage instructions
6. Link related skills
7. Update skills README
8. Commit new skill

创建的技能示例：

markdown

---
name: financial-calculator-builder
description: Build interactive financial calculators (NPV, IRR, ROI) with Plotly visualization, YAML configuration, and HTML reporting.
version: 1.0.0
category: development/finance
created_from: commit a1b2c3d4
pattern_source: npv_calculator implementation
---

Financial Calculator Builder

Create interactive financial analysis tools with visualization and reporting.

Quick Start

bash

undefined

bash

undefined

Create NPV calculator

/financial-calculator-builder npv

Create IRR calculator

/financial-calculator-builder irr

Custom calculator

/financial-calculator-builder custom --config config/calc.yaml

undefined

/financial-calculator-builder custom --config config/calc.yaml

undefined

Pattern

Define calculation in YAML config
Implement calculator class
Add Plotly visualization
Create bash execution wrapper
Generate HTML report
Write tests (TDD)

[... full skill documentation ...]

undefined

Define calculation in YAML config
Implement calculator class
Add Plotly visualization
Create bash execution wrapper
Generate HTML report
Write tests (TDD)

[... full skill documentation ...]

undefined

5. Skill Enhancement

5. 技能增强

Updates Existing Skills:

Enhancement Types:

markdown

1. **Version Updates**
   - New tool versions (Plotly 5.17 → 5.18)
   - Updated APIs
   - Deprecated method replacements

2. **Best Practice Improvements**
   - Better error handling
   - Performance optimizations
   - Security enhancements

3. **New Examples**
   - Additional use cases
   - Real-world implementations
   - Edge case handling

4. **Integration Points**
   - New tool integrations
   - Cross-skill workflows
   - Automation hooks

Example Enhancement:

markdown

undefined

更新现有技能：

增强类型：

markdown

1. **Version Updates**
   - New tool versions (Plotly 5.17 → 5.18)
   - Updated APIs
   - Deprecated method replacements

2. **Best Practice Improvements**
   - Better error handling
   - Performance optimizations
   - Security enhancements

3. **New Examples**
   - Additional use cases
   - Real-world implementations
   - Edge case handling

4. **Integration Points**
   - New tool integrations
   - Cross-skill workflows
   - Automation hooks

增强示例：

markdown

undefined

Enhancement: plotly-visualization skill

Commit: a1b2c3d4

Pattern Found: Multi-scenario NPV visualization

Enhancement Applied:

Added section "Financial Visualizations" to plotly-visualization skill:

New Example: Multi-Scenario Analysis

python

import plotly.graph_objects as go

Added section "Financial Visualizations" to plotly-visualization skill:

New Example: Multi-Scenario Analysis

python

import plotly.graph_objects as go

Create multi-scenario NPV plot

fig = go.Figure()

for scenario in scenarios: fig.add_trace(go.Scatter( x=scenario['years'], y=scenario['npv'], name=scenario['name'], mode='lines+markers' ))

fig.update_layout( title='NPV Analysis: Multiple Scenarios', xaxis_title='Year', yaxis_title='Net Present Value ($M)' )


**Version**: 1.2.0 → 1.3.0
**Reason**: Added financial visualization patterns

fig = go.Figure()

for scenario in scenarios: fig.add_trace(go.Scatter( x=scenario['years'], y=scenario['npv'], name=scenario['name'], mode='lines+markers' ))

fig.update_layout( title='NPV Analysis: Multiple Scenarios', xaxis_title='Year', yaxis_title='Net Present Value ($M)' )


**Version**: 1.2.0 → 1.3.0
**Reason**: Added financial visualization patterns

6. Knowledge Synthesis

6. 知识合成

Builds Organizational Knowledge:

Knowledge Base Structure:

.claude/knowledge/
├── patterns/
│   ├── financial-calculations.md
│   ├── interactive-visualization.md
│   └── data-pipelines.md
├── techniques/
│   ├── npv-calculation.md
│   ├── scenario-analysis.md
│   └── sensitivity-testing.md
└── lessons/
    ├── 2026-01-07-npv-calculator.md
    └── ...

Example Knowledge Entry:

markdown

undefined

构建组织知识：

知识库结构：

.claude/knowledge/
├── patterns/
│   ├── financial-calculations.md
│   ├── interactive-visualization.md
│   └── data-pipelines.md
├── techniques/
│   ├── npv-calculation.md
│   ├── scenario-analysis.md
│   └── sensitivity-testing.md
└── lessons/
    ├── 2026-01-07-npv-calculator.md
    └── ...

知识库条目示例：

markdown

undefined

Pattern: Interactive Financial Calculator

Discovered: 2026-01-07 (commit a1b2c3d4) Category: Financial Engineering Reusability: High

Problem

Need to perform financial calculations (NPV, IRR, ROI) with:

Multiple scenarios
Interactive parameter adjustment
Visual comparison
Exportable reports

Need to perform financial calculations (NPV, IRR, ROI) with:

Multiple scenarios
Interactive parameter adjustment
Visual comparison
Exportable reports

Solution Pattern

1. Configuration (YAML)

yaml

calculation:
  type: npv
  discount_rates: [0.05, 0.08, 0.10, 0.12]
  cash_flows: data/cash_flows.csv
  scenarios:
    - base
    - optimistic
    - pessimistic

yaml

calculation:
  type: npv
  discount_rates: [0.05, 0.08, 0.10, 0.12]
  cash_flows: data/cash_flows.csv
  scenarios:
    - base
    - optimistic
    - pessimistic

2. Calculator Module (Python)

python

class NPVCalculator:
    def calculate(self, cash_flows, discount_rate):
        return np.npv(discount_rate, cash_flows)

    def multi_scenario(self, scenarios, rates):
        results = {}
        for scenario in scenarios:
            for rate in rates:
                npv = self.calculate(scenario.cash_flows, rate)
                results[(scenario.name, rate)] = npv
        return results

python

class NPVCalculator:
    def calculate(self, cash_flows, discount_rate):
        return np.npv(discount_rate, cash_flows)

    def multi_scenario(self, scenarios, rates):
        results = {}
        for scenario in scenarios:
            for rate in rates:
                npv = self.calculate(scenario.cash_flows, rate)
                results[(scenario.name, rate)] = npv
        return results

3. Visualization (Plotly)

Interactive multi-scenario comparison with hover tooltips.

4. Execution (Bash)

bash

./scripts/run_npv_analysis.sh config/npv.yaml

bash

./scripts/run_npv_analysis.sh config/npv.yaml

Lessons Learned

YAML configuration makes calculators flexible
Plotly enables interactive scenario exploration
Modular design allows easy extension to other metrics
TDD catches calculation errors early

YAML configuration makes calculators flexible
Plotly enables interactive scenario exploration
Modular design allows easy extension to other metrics
TDD catches calculation errors early

Applications

✅ NPV Calculator (implemented)
🔲 IRR Calculator (recommended)
🔲 ROI Calculator (recommended)
🔲 Payback Period Calculator (recommended)
🔲 Sensitivity Analysis Tool (high value)

✅ NPV Calculator (implemented)
🔲 IRR Calculator (recommended)
🔲 ROI Calculator (recommended)
🔲 Payback Period Calculator (recommended)
🔲 Sensitivity Analysis Tool (high value)

Related Skills

financial-calculator-builder (created)
plotly-visualization (enhanced)
yaml-workflow-executor (existing)

undefined

financial-calculator-builder (created)
plotly-visualization (enhanced)
yaml-workflow-executor (existing)

undefined

Learning Log

学习日志

Maintains Continuous Learning Record:

Log Format:

markdown

undefined

维护持续学习记录：

日志格式：

markdown

undefined

Skill Learning Log

2026-01-07

Commit: a1b2c3d4 - NPV Calculator Implementation

Patterns Extracted: 3

Interactive financial calculator
Multi-scenario analysis
YAML-driven calculations

Skills Created: 1

financial-calculator-builder (v1.0.0)

Skills Enhanced: 2

plotly-visualization (v1.2.0 → v1.3.0)
- Added financial visualization examples
yaml-workflow-executor (v2.1.0 → v2.1.1)
- Added financial config examples

Knowledge Added: 1

patterns/interactive-financial-calculator.md

Reusability Score: 95/100 Time Savings: ~2 hours per future use Learning Value: High - establishes financial tool pattern

Patterns Extracted: 3

Interactive financial calculator
Multi-scenario analysis
YAML-driven calculations

Skills Created: 1

financial-calculator-builder (v1.0.0)

Skills Enhanced: 2

plotly-visualization (v1.2.0 → v1.3.0)
- Added financial visualization examples
yaml-workflow-executor (v2.1.0 → v2.1.1)
- Added financial config examples

Knowledge Added: 1

patterns/interactive-financial-calculator.md

Reusability Score: 95/100 Time Savings: ~2 hours per future use Learning Value: High - establishes financial tool pattern

2026-01-06

Commit: e5f6g7h8 - Marine Safety Data Processor

Patterns Extracted: 2

CSV data validation
Safety incident categorization

Skills Enhanced: 1

data-pipeline-processor (v3.0.0 → v3.1.0)
- Added safety data validation patterns

Knowledge Added: 1

techniques/safety-data-validation.md

Reusability Score: 75/100 Time Savings: ~1 hour per future use Learning Value: Medium - domain-specific but valuable

undefined

Patterns Extracted: 2

CSV data validation
Safety incident categorization

Skills Enhanced: 1

data-pipeline-processor (v3.0.0 → v3.1.0)
- Added safety data validation patterns

Knowledge Added: 1

techniques/safety-data-validation.md

Reusability Score: 75/100 Time Savings: ~1 hour per future use Learning Value: Medium - domain-specific but valuable

undefined

Execution Checklist

执行检查清单

Hook Integration

钩子集成

Post-Commit Hook

提交后钩子

Hook Configuration:

bash

undefined

钩子配置：

bash

undefined

.claude/hooks/post-commit.sh

#!/bin/bash

Auto-execute skill learning after commits

REPO_PATH="$(pwd)" SKILL_PATH="${HOME}/.claude/skills/workspace-hub/skill-learner"

Allow bypassing skill learning

if [ "${SKIP_SKILL_LEARNING:-0}" = "1" ]; then echo "Skill learning skipped (SKIP_SKILL_LEARNING=1)" exit 0 fi

Only run on significant commits

LINES_CHANGED=$(git diff HEAD^ HEAD --shortstat | grep -oE '[0-9]+ insertion' | grep -oE '[0-9]+' || echo 0)

if [ "$LINES_CHANGED" -lt 50 ]; then echo "Small commit (<50 lines), skipping skill learning" exit 0 fi

LINES_CHANGED=$(git diff HEAD^ HEAD --shortstat | grep -oE '[0-9]+ insertion' | grep -oE '[0-9]+' || echo 0)

if [ "$LINES_CHANGED" -lt 50 ]; then echo "Small commit (<50 lines), skipping skill learning" exit 0 fi

Run skill learning

echo "Analyzing commit for learning opportunities..." "$SKILL_PATH/analyze_commit.sh" "$REPO_PATH"

exit 0


**Trigger Conditions:**
- After git commit (>50 lines changed)
- After task completion
- After feature merges
- Manual trigger via `/skill-learner`

**Bypass Hook:**
```bash

echo "Analyzing commit for learning opportunities..." "$SKILL_PATH/analyze_commit.sh" "$REPO_PATH"

exit 0


**触发条件：**
- git提交后（修改行数>50）
- 任务完成后
- 功能合并后
- 通过`/skill-learner`手动触发

**跳过钩子：**
```bash

Skip learning for trivial commits

SKIP_SKILL_LEARNING=1 git commit -m "Fix typo"

Or disable temporarily

mv .claude/hooks/post-commit.sh .claude/hooks/post-commit.sh.disabled

undefined

mv .claude/hooks/post-commit.sh .claude/hooks/post-commit.sh.disabled

undefined

Automation Scripts

自动化脚本

1. Commit Analyzer

1. 提交分析器

Location:

analyze_commit.sh

bash

#!/bin/bash

位置：

analyze_commit.sh

bash

#!/bin/bash

Analyze recent commit for learning opportunities

REPO_PATH="${1:-.}" COMMIT_HASH="${2:-HEAD}"

Extract commit info

analyze_commit() { echo "Analyzing commit: $COMMIT_HASH"

# Get commit metadata
local msg=$(git log -1 --pretty=%B $COMMIT_HASH)
local author=$(git log -1 --pretty=%an $COMMIT_HASH)
local date=$(git log -1 --pretty=%ai $COMMIT_HASH)

# Get changed files
local files=$(git diff-tree --no-commit-id --name-only -r $COMMIT_HASH)

# Analyze patterns
extract_patterns "$files"

# Make skill decisions
decide_skill_actions

# Generate report
generate_learning_report

}

analyze_commit() { echo "Analyzing commit: $COMMIT_HASH"

# Get commit metadata
local msg=$(git log -1 --pretty=%B $COMMIT_HASH)
local author=$(git log -1 --pretty=%an $COMMIT_HASH)
local date=$(git log -1 --pretty=%ai $COMMIT_HASH)

# Get changed files
local files=$(git diff-tree --no-commit-id --name-only -r $COMMIT_HASH)

# Analyze patterns
extract_patterns "$files"

# Make skill decisions
decide_skill_actions

# Generate report
generate_learning_report

}

Extract reusable patterns

extract_patterns() { local files="$1"

# Check for workflow patterns
if echo "$files" | grep -q "config/input/.*\.yaml"; then
    echo "Pattern: YAML-driven workflow"
fi

if echo "$files" | grep -q "scripts/.*\.sh"; then
    echo "Pattern: Bash execution script"
fi

# Check for code patterns
if git diff $COMMIT_HASH^ $COMMIT_HASH | grep -q "import plotly"; then
    echo "Pattern: Plotly visualization"
fi

# More pattern detection...

}

extract_patterns() { local files="$1"

# Check for workflow patterns
if echo "$files" | grep -q "config/input/.*\.yaml"; then
    echo "Pattern: YAML-driven workflow"
fi

if echo "$files" | grep -q "scripts/.*\.sh"; then
    echo "Pattern: Bash execution script"
fi

# Check for code patterns
if git diff $COMMIT_HASH^ $COMMIT_HASH | grep -q "import plotly"; then
    echo "Pattern: Plotly visualization"
fi

# More pattern detection...

}

Decide whether to create/enhance skills

decide_skill_actions() { # Reusability scoring local score=0

# Check commit history for similar patterns
local pattern_count=$(git log --all --grep="similar pattern" | wc -l)
if [ $pattern_count -ge 3 ]; then
    ((score += 30))
fi

# Check code complexity
local lines_added=$(git diff $COMMIT_HASH^ $COMMIT_HASH --shortstat | grep -oE '[0-9]+ insertion' | grep -oE '[0-9]+')
if [ $lines_added -gt 100 ]; then
    ((score += 20))
fi

# Decision
if [ $score -ge 70 ]; then
    echo "Recommendation: CREATE NEW SKILL"
elif [ $score -ge 40 ]; then
    echo "Recommendation: ENHANCE EXISTING SKILL"
else
    echo "Recommendation: SKIP (low reusability)"
fi

}

decide_skill_actions() { # Reusability scoring local score=0

# Check commit history for similar patterns
local pattern_count=$(git log --all --grep="similar pattern" | wc -l)
if [ $pattern_count -ge 3 ]; then
    ((score += 30))
fi

# Check code complexity
local lines_added=$(git diff $COMMIT_HASH^ $COMMIT_HASH --shortstat | grep -oE '[0-9]+ insertion' | grep -oE '[0-9]+')
if [ $lines_added -gt 100 ]; then
    ((score += 20))
fi

# Decision
if [ $score -ge 70 ]; then
    echo "Recommendation: CREATE NEW SKILL"
elif [ $score -ge 40 ]; then
    echo "Recommendation: ENHANCE EXISTING SKILL"
else
    echo "Recommendation: SKIP (low reusability)"
fi

}

Generate learning report

generate_learning_report() { local report_file=".claude/learning-reports/$(date +%Y-%m-%d)-$COMMIT_HASH.md" mkdir -p "$(dirname "$report_file")"

{
    echo "# Learning Report"
    echo "**Commit**: $COMMIT_HASH"
    echo "**Date**: $(date)"
    echo ""
    echo "## Patterns Extracted"
    # ... pattern details ...
    echo ""
    echo "## Recommendations"
    # ... recommendations ...
} > "$report_file"

echo "Learning report saved: $report_file"

}

main "$@"

undefined

generate_learning_report() { local report_file=".claude/learning-reports/$(date +%Y-%m-%d)-$COMMIT_HASH.md" mkdir -p "$(dirname "$report_file")"

{
    echo "# Learning Report"
    echo "**Commit**: $COMMIT_HASH"
    echo "**Date**: $(date)"
    echo ""
    echo "## Patterns Extracted"
    # ... pattern details ...
    echo ""
    echo "## Recommendations"
    # ... recommendations ...
} > "$report_file"

echo "Learning report saved: $report_file"

}

main "$@"

undefined

2. Skill Creator (from Pattern)

2. 从模式创建技能

Location:

create_skill_from_pattern.sh

bash

#!/bin/bash

位置：

create_skill_from_pattern.sh

bash

#!/bin/bash

Create new skill from extracted pattern

PATTERN_NAME="$1" CATEGORY="${2:-development}" SOURCE_COMMIT="$3"

Generate skill from template

generate_skill() { local skill_dir="${HOME}/.claude/skills/${CATEGORY}/${PATTERN_NAME}" mkdir -p "$skill_dir"

# Create SKILL.md
cat > "${skill_dir}/SKILL.md" << EOF

name: ${PATTERN_NAME} description: Auto-generated skill from commit pattern version: 1.0.0 category: ${CATEGORY} created_from: ${SOURCE_COMMIT}

generate_skill() { local skill_dir="${HOME}/.claude/skills/${CATEGORY}/${PATTERN_NAME}" mkdir -p "$skill_dir"

# Create SKILL.md
cat > "${skill_dir}/SKILL.md" << EOF

name: ${PATTERN_NAME} description: Auto-generated skill from commit pattern version: 1.0.0 category: ${CATEGORY} created_from: ${SOURCE_COMMIT}

${PATTERN_NAME^} Skill

Auto-generated from repeated commit pattern

Pattern Source

Commit: ${SOURCE_COMMIT} Extracted: $(date)

[... skill template ...] EOF

echo "Skill created: $skill_dir"

}

generate_skill

undefined

Commit: ${SOURCE_COMMIT} Extracted: $(date)

[... skill template ...] EOF

echo "Skill created: $skill_dir"

}

generate_skill

undefined

3. Bulk Learning Analysis

3. 批量学习分析

Analyzes Recent Commit History:

bash

#!/bin/bash

分析近期提交历史：

bash

#!/bin/bash

Analyze last N commits for learning opportunities

REPO_PATH="${1:-.}" COMMIT_COUNT="${2:-10}"

echo "Analyzing last $COMMIT_COUNT commits..."

for commit in $(git log -n $COMMIT_COUNT --pretty=%H); do echo "Commit: $commit" ./analyze_commit.sh "$REPO_PATH" "$commit" echo "" done

REPO_PATH="${1:-.}" COMMIT_COUNT="${2:-10}"

echo "Analyzing last $COMMIT_COUNT commits..."

for commit in $(git log -n $COMMIT_COUNT --pretty=%H); do echo "Commit: $commit" ./analyze_commit.sh "$REPO_PATH" "$commit" echo "" done

Generate aggregate report

echo "Generating aggregate learning report..." ./generate_aggregate_report.sh

undefined

echo "Generating aggregate learning report..." ./generate_aggregate_report.sh

undefined

Error Handling

错误处理

No Patterns Found

未找到模式

ℹ️ Info: No reusable patterns detected in commit

Reason: Commit too small or repository-specific changes

Action: No skills created, logged for future pattern detection

ℹ️ Info: No reusable patterns detected in commit

Reason: Commit too small or repository-specific changes

Action: No skills created, logged for future pattern detection

Skill Already Exists

技能已存在

⚠️ Warning: Skill 'financial-calculator-builder' already exists

Action Options:
1. Enhance existing skill (recommended)
2. Create variant skill (e.g., 'financial-calculator-builder-v2')
3. Skip creation

Recommendation: ENHANCE

⚠️ Warning: Skill 'financial-calculator-builder' already exists

Action Options:
1. Enhance existing skill (recommended)
2. Create variant skill (e.g., 'financial-calculator-builder-v2')
3. Skip creation

Recommendation: ENHANCE

Invalid Pattern

无效模式

❌ Error: Pattern extraction failed

Reason: Unable to identify coherent workflow or technique

Action: Manual review required - see .claude/learning-reports/error-*.md

❌ Error: Pattern extraction failed

Reason: Unable to identify coherent workflow or technique

Action: Manual review required - see .claude/learning-reports/error-*.md

Skills Directory Not Found

技能目录未找到

⚠️ Warning: Skills directory not accessible

Location Expected: ~/.claude/skills/ or .claude/skills/

Action: Create skills directory or configure SKILLS_PATH

⚠️ Warning: Skills directory not accessible

Location Expected: ~/.claude/skills/ or .claude/skills/

Action: Create skills directory or configure SKILLS_PATH

Metrics & Success Criteria

指标与成功标准

Learning Metrics

学习指标

Metric	Target	Measurement
Patterns Extracted/Week	5-10	Count from learning log
Skills Created/Month	2-5	New skills in directory
Skills Enhanced/Month	5-10	Version bumps recorded
Reusability Score Avg	> 70	Average across patterns
Time Savings	> 10 hrs/month	Estimated from usage

指标	目标	测量方式
每周提取的模式数量	5-10	从学习日志统计
每月创建的技能数量	2-5	目录中的新技能
每月增强的技能数量	5-10	�记录的版本更新
平均可复用性得分	> 70	所有模式的平均值
时间节省	> 10小时/月	根据使用情况估算

Quality Metrics

质量指标

Metric	Target	Measurement
Pattern Accuracy	> 90%	Correct pattern identification
Skill Usefulness	> 80%	Skills used after creation
Enhancement Value	> 70%	Enhanced skills improve workflow
False Positives	< 10%	Incorrect skill creation rate

指标	目标	测量方式
模式准确率	> 90%	正确识别模式的比例
技能实用性	> 80%	创建后被使用的技能比例
增强价值	> 70%	增强后的技能对工作流的提升程度
误报率	< 10%	错误创建技能的比例

Adoption Metrics

采用指标

Metric	Target	Current
Hook Installation	100% repos	Track per repo
Learning Log Size	Growing	Lines in log
Knowledge Base Growth	+10 entries/month	Count entries
Skill Library Size	+3-5 skills/month	Total skills

指标	目标	当前值
钩子安装率	100%仓库	按仓库统计
学习日志规模	持续增长	日志行数
知识库增长	+10条目/月	条目数量
技能库规模	+3-5技能/月	总技能数量

Integration Points

集成点

With Session Start Routine

与会话启动流程集成

bash

undefined

bash

undefined

Review recent learning at session start

/session-start-routine

Includes:

- Recent patterns extracted

- Skills created/enhanced

- Learning opportunities

undefined

undefined

With Skill Creator

与技能创建器集成

bash

undefined

bash

undefined

Manual skill creation uses patterns

/skill-creator new-feature

Skill learner provides:

- Similar existing patterns

- Related skills

- Best practices from learning log

undefined

undefined

With Repo Sync

与仓库同步集成

bash

undefined

bash

undefined

Learning across all repos

./bulk_learning_analysis.sh

Aggregate patterns from entire workspace

undefined

undefined

Best Practices

最佳实践

1. Review Learning Log Regularly

1. 定期查看学习日志

bash

undefined

bash

undefined

Weekly review

cat .claude/skill-learning-log.md | head -100

Identify high-value patterns

grep "Reusability Score: 9" .claude/skill-learning-log.md

undefined

grep "Reusability Score: 9" .claude/skill-learning-log.md

undefined

2. Refine Patterns

2. 优化模式

bash

undefined

bash

undefined

If pattern is too specific, generalize

If pattern is too generic, specialize

Document edge cases

undefined

undefined

3. Link Related Knowledge

3. 关联相关知识

bash

undefined

bash

undefined

Cross-reference skills

Update related documentation

Share patterns across teams

undefined

undefined

4. Validate Skill Usefulness

4. 验证技能实用性

bash

undefined

bash

undefined

Track skill usage

Deprecate unused skills

Enhance frequently-used skills

undefined

undefined

Related Skills

References

参考资料

Skill Template
Pattern Library
Learning Framework

Skill Template
Pattern Library
Learning Framework

Version History

版本历史

1.0.0 (2026-01-07): Initial release - post-commit skill learning with pattern extraction, skill creation/enhancement, knowledge synthesis, auto-hook integration, learning log, metrics tracking, and continual learning capabilities

1.0.0 (2026-01-07): 初始版本 - 提交后技能学习，包含模式提取、技能创建/增强、知识合成、自动钩子集成、学习日志、指标跟踪和持续学习功能

skill-learner

Original

Translation

Skill Learner

Skill Learner

Quick Start

快速开始

Manual trigger after commit

Manual trigger after commit

Auto-triggers after:

Auto-triggers after:

- git commit (via post-commit hook)

- git commit (via post-commit hook)

- Task completion

- Task completion

- Feature implementation

- Feature implementation

Review learning

Review learning

When to Use

适用场景

Prerequisites

前置条件

Overview

概述

What It Analyzes

分析内容

Output

输出内容

Core Operations

核心操作

1. Commit Analysis

1. 提交分析

Extract commit metadata

Extract commit metadata

Get changed files

Get changed files

Get diff content

Get diff content

Commit Analysis

Commit Analysis

Files Changed (12)

Files Changed (12)

Code Additions

Code Additions

Technologies Used

Technologies Used

2. Pattern Extraction

2. 模式提取

Pattern: Interactive Financial Calculator

Pattern: Interactive Financial Calculator

Problem Solved

Problem Solved

Solution Pattern

Solution Pattern

Reusability Score: 95/100

Reusability Score: 95/100

3. Skill Identification

3. 技能识别

Skill Decision: NPV Calculator

Skill Decision: NPV Calculator

Analysis

Analysis

Recommendation: CREATE NEW SKILL

Recommendation: CREATE NEW SKILL

4. Skill Creation

4. 技能创建

Financial Calculator Builder

Financial Calculator Builder

Quick Start

Quick Start

Create NPV calculator

Create NPV calculator

Create IRR calculator

Create IRR calculator

Custom calculator

Custom calculator

Pattern

Pattern

5. Skill Enhancement