Pattern Extraction

Pattern Extraction Skill

When to Use This Skill

Use pattern extraction when you need to:

• Analyze existing infrastructure to create reusable templates
• Convert hardcoded values into configurable template variables
• Identify configuration patterns across multiple similar files
• Detect technology stacks and their conventions
• Extract naming conventions and structural patterns
• Generate template metadata from existing code
• Create variable schemas from inferred types and constraints

Perfect for:

• Converting existing IaC to templates
• Building template libraries from production code
• Standardizing infrastructure patterns
• Automating template generation
• Identifying refactoring opportunities

Core Capabilities

1. Technology Stack Detection

Automatically identify:

• IaC Frameworks: Terraform, Pulumi, CloudFormation, ARM, Bicep
• Cloud Providers: AWS, Azure, GCP, multi-cloud patterns
• Service Types: Kubernetes, Docker, serverless, containers
• Configuration Formats: YAML, JSON, HCL, TOML
• Build Systems: Helm, Kustomize, Jsonnet
• CI/CD Platforms: GitHub Actions, GitLab CI, Jenkins, Harness

2. Configuration Pattern Extraction

Extract patterns from:

• Resource definitions and relationships
• Environment-specific configurations
• Naming conventions and tagging strategies
• Security policies and compliance rules
• Network topologies and architectures
• Deployment strategies and workflows

3. Variable Identification

Intelligent detection of:

• Hardcoded Values: Strings, numbers, booleans that should be variables
• Repeated Values: Values appearing multiple times across files
• Environment Indicators: dev, staging, prod patterns
• Naming Patterns: Prefixes, suffixes, delimiters
• Secret Patterns: API keys, passwords, tokens (flag for security)
• Configuration Schemas: Type inference from usage

4. Structure Analysis

Analyze:

• File organization and directory structure
• Module boundaries and dependencies
• Resource hierarchies and relationships
• Configuration inheritance patterns
• Composition and reuse strategies
• Template inclusion patterns

Pattern Detection Matrix

Variable Inference Rules

Before/After Transformation Examples

Example 1: Resource Names

Before:

resource "aws_instance" "web_server" {
  ami           = "ami-0c55b159cbfafe1f0"
  instance_type = "t3.medium"

  tags = {
    Name        = "myapp-web-server-prod"
    Environment = "production"
    Project     = "myapp"
  }
}

After:

resource "aws_instance" "web_server" {
  ami           = "{{ ami_id }}"
  instance_type = "{{ instance_type }}"

  tags = {
    Name        = "{{ project_name }}-web-server-{{ environment }}"
    Environment = "{{ environment }}"
    Project     = "{{ project_name }}"
  }
}

Extracted Variables:

variables:
  ami_id:
    type: string
    description: "AMI ID for the EC2 instance"
    default: "ami-0c55b159cbfafe1f0"
    pattern: "^ami-[a-f0-9]{17}$"

  instance_type:
    type: string
    description: "EC2 instance type"
    default: "t3.medium"
    allowed_values: ["t3.micro", "t3.small", "t3.medium", "t3.large"]

  project_name:
    type: string
    description: "Project identifier used in resource naming"
    default: "myapp"
    pattern: "^[a-z][a-z0-9-]{2,30}$"

  environment:
    type: string
    description: "Deployment environment"
    default: "production"
    allowed_values: ["dev", "staging", "production"]

Example 2: Kubernetes Deployment

Before:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
  namespace: production
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
        version: "1.21.0"
    spec:
      containers:
      - name: nginx
        image: nginx:1.21.0
        ports:
        - containerPort: 80
        resources:
          requests:
            memory: "128Mi"
            cpu: "250m"
          limits:
            memory: "256Mi"
            cpu: "500m"

After:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: {{ app_name }}-deployment
  namespace: {{ namespace }}
spec:
  replicas: {{ replica_count }}
  selector:
    matchLabels:
      app: {{ app_name }}
  template:
    metadata:
      labels:
        app: {{ app_name }}
        version: "{{ app_version }}"
    spec:
      containers:
      - name: {{ app_name }}
        image: {{ container_image }}:{{ app_version }}
        ports:
        - containerPort: {{ container_port }}
        resources:
          requests:
            memory: "{{ memory_request }}"
            cpu: "{{ cpu_request }}"
          limits:
            memory: "{{ memory_limit }}"
            cpu: "{{ cpu_limit }}"

Extracted Variables:

variables:
  app_name:
    type: string
    description: "Application name"
    default: "nginx"

  namespace:
    type: string
    description: "Kubernetes namespace"
    default: "production"

  replica_count:
    type: integer
    description: "Number of pod replicas"
    default: 3
    min: 1
    max: 10

  app_version:
    type: string
    description: "Application version"
    default: "1.21.0"
    pattern: "^\\d+\\.\\d+\\.\\d+$"

  container_image:
    type: string
    description: "Container image name"
    default: "nginx"

  container_port:
    type: integer
    description: "Container port"
    default: 80

  memory_request:
    type: string
    description: "Memory request"
    default: "128Mi"

  memory_limit:
    type: string
    description: "Memory limit"
    default: "256Mi"

  cpu_request:
    type: string
    description: "CPU request"
    default: "250m"

  cpu_limit:
    type: string
    description: "CPU limit"
    default: "500m"

Example 3: Azure Bicep

Before:

resource storageAccount 'Microsoft.Storage/storageAccounts@2021-04-01' = {
  name: 'mystorageacct12345'
  location: 'westeurope'
  sku: {
    name: 'Standard_LRS'
  }
  kind: 'StorageV2'
  properties: {
    accessTier: 'Hot'
    minimumTlsVersion: 'TLS1_2'
    supportsHttpsTrafficOnly: true
    allowBlobPublicAccess: false
  }
  tags: {
    environment: 'production'
    costCenter: 'engineering'
    project: 'platform'
  }
}

After:

resource storageAccount 'Microsoft.Storage/storageAccounts@2021-04-01' = {
  name: '{{ storage_account_name }}'
  location: '{{ location }}'
  sku: {
    name: '{{ sku_name }}'
  }
  kind: 'StorageV2'
  properties: {
    accessTier: '{{ access_tier }}'
    minimumTlsVersion: '{{ min_tls_version }}'
    supportsHttpsTrafficOnly: {{ https_only }}
    allowBlobPublicAccess: {{ allow_public_access }}
  }
  tags: {
    environment: '{{ environment }}'
    costCenter: '{{ cost_center }}'
    project: '{{ project_name }}'
  }
}

Extracted Variables:

variables:
  storage_account_name:
    type: string
    description: "Storage account name (globally unique)"
    default: "mystorageacct12345"
    pattern: "^[a-z0-9]{3,24}$"

  location:
    type: string
    description: "Azure region"
    default: "westeurope"
    allowed_values: ["westeurope", "northeurope", "eastus", "westus"]

  sku_name:
    type: string
    description: "Storage account SKU"
    default: "Standard_LRS"
    allowed_values: ["Standard_LRS", "Standard_GRS", "Premium_LRS"]

  access_tier:
    type: string
    description: "Storage access tier"
    default: "Hot"
    allowed_values: ["Hot", "Cool"]

  min_tls_version:
    type: string
    description: "Minimum TLS version"
    default: "TLS1_2"

  https_only:
    type: boolean
    description: "Require HTTPS traffic only"
    default: true

  allow_public_access:
    type: boolean
    description: "Allow public blob access"
    default: false

  environment:
    type: string
    description: "Environment name"
    default: "production"

  cost_center:
    type: string
    description: "Cost center for billing"
    default: "engineering"

  project_name:
    type: string
    description: "Project name"
    default: "platform"

Decision Tree

Pattern classification flowchart:

START: Analyze value/pattern
    |
    v
┌───────────────────────────────────┐
│ Is it repeated across files?      │
└─────┬─────────────────────┬───────┘
      │ YES                 │ NO
      v                     v
┌─────────────┐      ┌──────────────┐
│ Global Var  │      │ Check Usage  │
└─────────────┘      └──────┬───────┘
                            │
                            v
                     ┌──────────────────────┐
                     │ Used in conditionals? │
                     └──┬─────────────────┬──┘
                        │ YES             │ NO
                        v                 v
                 ┌──────────┐      ┌──────────────┐
                 │ Feature  │      │ Check Type   │
                 │ Flag     │      └──────┬───────┘
                 └──────────┘             │
                                          v
                                   ┌──────────────────────┐
                                   │ Contains 'env'?      │
                                   └──┬───────────────┬───┘
                                      │ YES           │ NO
                                      v               v
                                ┌───────────┐   ┌──────────────┐
                                │ Env Var   │   │ Check Format │
                                └───────────┘   └──────┬───────┘
                                                       │
                                                       v
                                                ┌──────────────────┐
                                                │ Cloud resource?  │
                                                └──┬───────────┬───┘
                                                   │ YES       │ NO
                                                   v           v
                                            ┌──────────┐  ┌─────────┐
                                            │ Resource │  │ Generic │
                                            │ ID       │  │ Config  │
                                            └──────────┘  └─────────┘

Examples

Example 1: Extract Terraform AWS Pattern

Input:

extract patterns from ./terraform/aws/ec2-instances.tf

Analysis:

Analyzing: terraform/aws/ec2-instances.tf
├── Technology: Terraform (AWS Provider)
├── Resources Found: 3 (aws_instance, aws_security_group, aws_eip)
├── Variables Identified: 12
└── Patterns Detected:
    ├── Naming Convention: {project}-{resource}-{env}
    ├── Tag Strategy: Environment, Project, ManagedBy
    └── Configuration Reuse: 85% similarity across instances

Extracted Template:

# Template: aws-ec2-instance.tf.tmpl
variable "project_name" {
  type        = string
  description = "Project identifier"
}

variable "environment" {
  type        = string
  description = "Deployment environment"
  validation {
    condition     = contains(["dev", "staging", "prod"], var.environment)
    error_message = "Environment must be dev, staging, or prod."
  }
}

variable "instance_config" {
  type = object({
    ami           = string
    instance_type = string
    key_name      = string
    subnet_id     = string
  })
  description = "EC2 instance configuration"
}

resource "aws_instance" "main" {
  ami           = var.instance_config.ami
  instance_type = var.instance_config.instance_type
  key_name      = var.instance_config.key_name
  subnet_id     = var.instance_config.subnet_id

  vpc_security_group_ids = [aws_security_group.main.id]

  tags = {
    Name        = "${var.project_name}-instance-${var.environment}"
    Environment = var.environment
    Project     = var.project_name
    ManagedBy   = "Terraform"
  }
}

resource "aws_security_group" "main" {
  name        = "${var.project_name}-sg-${var.environment}"
  description = "Security group for ${var.project_name} in ${var.environment}"

  tags = {
    Name        = "${var.project_name}-sg-${var.environment}"
    Environment = var.environment
    Project     = var.project_name
  }
}

Example 2: Extract Kubernetes Pattern

Input:

extract patterns from ./k8s/deployments/ --type kubernetes

Analysis:

Analyzing: k8s/deployments/ (15 files)
├── Technology: Kubernetes (v1.24+)
├── Resources Found: 45 total
│   ├── Deployments: 15
│   ├── Services: 12
│   ├── ConfigMaps: 10
│   └── Ingresses: 8
├── Common Patterns:
│   ├── Label Strategy: app, version, environment
│   ├── Resource Requests: 90% use standard sizes
│   ├── Probes: 100% have health checks
│   └── Image Pattern: registry/org/image:tag
└── Variable Candidates: 28 identified

Extracted Template:

# Template: kubernetes-microservice.yaml.tmpl
apiVersion: apps/v1
kind: Deployment
metadata:
  name: {{ service_name }}-deployment
  namespace: {{ namespace }}
  labels:
    app: {{ service_name }}
    version: {{ version }}
    environment: {{ environment }}
spec:
  replicas: {{ replica_count }}
  selector:
    matchLabels:
      app: {{ service_name }}
  template:
    metadata:
      labels:
        app: {{ service_name }}
        version: {{ version }}
        environment: {{ environment }}
    spec:
      containers:
      - name: {{ service_name }}
        image: {{ image_registry }}/{{ organization }}/{{ service_name }}:{{ version }}
        ports:
        - containerPort: {{ container_port }}
          protocol: TCP
        env:
        {{#each environment_variables}}
        - name: {{ name }}
          value: "{{ value }}"
        {{/each}}
        resources:
          requests:
            memory: {{ memory_request }}
            cpu: {{ cpu_request }}
          limits:
            memory: {{ memory_limit }}
            cpu: {{ cpu_limit }}
        livenessProbe:
          httpGet:
            path: {{ health_check_path }}
            port: {{ container_port }}
          initialDelaySeconds: {{ liveness_initial_delay }}
          periodSeconds: {{ liveness_period }}
        readinessProbe:
          httpGet:
            path: {{ readiness_check_path }}
            port: {{ container_port }}
          initialDelaySeconds: {{ readiness_initial_delay }}
          periodSeconds: {{ readiness_period }}
---
apiVersion: v1
kind: Service
metadata:
  name: {{ service_name }}-service
  namespace: {{ namespace }}
  labels:
    app: {{ service_name }}
spec:
  type: {{ service_type }}
  ports:
  - port: {{ service_port }}
    targetPort: {{ container_port }}
    protocol: TCP
  selector:
    app: {{ service_name }}

Example 3: Multi-File Pattern Extraction

Input:

extract patterns from ./infrastructure/ --recursive --consolidate

Analysis:

Analyzing: infrastructure/ (recursive)
├── Files Scanned: 47
├── Technologies Detected:
│   ├── Terraform (AWS): 23 files
│   ├── Kubernetes: 15 files
│   ├── Helm Charts: 9 files
│   └── Docker Compose: 2 files (excluded - different pattern)
├── Cross-Cutting Patterns:
│   ├── Environment Strategy: 3-tier (dev/staging/prod)
│   ├── Tagging: 100% compliance with org policy
│   ├── Naming: Consistent kebab-case with env suffix
│   └── Secrets: HashiCorp Vault references
└── Template Opportunities:
    ├── AWS Lambda Function: 8 similar resources
    ├── RDS Database: 5 similar resources
    ├── Kubernetes Service: 12 similar resources
    └── ALB Configuration: 6 similar resources

Output:

Generated Templates:
├── templates/
│   ├── aws-lambda-function.tf.tmpl (consolidated from 8 files)
│   ├── aws-rds-instance.tf.tmpl (consolidated from 5 files)
│   ├── kubernetes-service.yaml.tmpl (consolidated from 12 files)
│   └── aws-alb.tf.tmpl (consolidated from 6 files)
└── variables/
    ├── common.yaml (shared across all templates)
    ├── aws-specific.yaml (AWS provider variables)
    └── kubernetes-specific.yaml (K8s variables)

Variable Reuse Analysis:
├── Shared Variables: 15 (45% reuse)
├── Template-Specific: 18 (55% unique)
└── Potential Consolidation: 3 variables can be merged

Best Practices

1. Incremental Extraction

Start with a single file or small directory to refine patterns before scaling:

# Start small
extract patterns from ./terraform/main.tf

# Validate and adjust
review template ./templates/main.tf.tmpl

# Scale up
extract patterns from ./terraform/ --recursive

2. Variable Naming Conventions

Follow consistent naming:

• Use snake_case for variables
• Prefix with context: aws_, k8s_, azure_
• Suffix with type: _count, _enabled, _config
• Be descriptive: instance_type not type

3. Type Inference Priority

1. Explicit types from existing variable definitions
2. Usage patterns (e.g., used in math → number)
3. Value format (e.g., "true"/"false" → boolean)
4. Defaults to string if ambiguous

4. Security-Sensitive Patterns

Always flag potential secrets:

# Good - flagged for security review
api_key:
  type: string
  sensitive: true
  description: "API key for external service"
  default: null  # Force explicit value

5. Validation Rules

Add validation for critical variables:

environment:
  type: string
  allowed_values: ["dev", "staging", "prod"]

cidr_block:
  type: string
  pattern: "^\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}\\.\\d{1,3}/\\d{1,2}$"

replica_count:
  type: integer
  min: 1
  max: 100

6. Documentation Generation

Auto-generate docs from extracted patterns:

# Generated Template: aws-ec2-instance

## Description
Extracted from 8 similar EC2 instance definitions.
Pattern confidence: 95%

## Variables
| Name | Type | Required | Default | Description |
|------|------|----------|---------|-------------|
| instance_type | string | Yes | t3.medium | EC2 instance type |
| environment | string | Yes | - | Deployment environment |

## Usage
terraform apply -var="instance_type=t3.large" -var="environment=prod"

7. Pattern Confidence Scoring

Rate extraction confidence:

• High (90-100%): Identical patterns across files
• Medium (70-89%): Similar with minor variations
• Low (<70%): Significant differences, manual review needed

8. Iterative Refinement

# Extract initial patterns
extract patterns from ./infra/ --output ./templates/v1/

# Review and refine
review templates ./templates/v1/ --suggest-improvements

# Re-extract with refinements
extract patterns from ./infra/ --output ./templates/v2/ \
  --naming-convention kebab-case \
  --tag-strategy company-standard

Related Skills

• [[template-generation]] - Generate templates from extracted patterns
• [[variable-schema-design]] - Design robust variable schemas
• [[template-validation]] - Validate generated templates
• [[naming-convention-analyzer]] - Analyze and enforce naming conventions
• [[configuration-consolidation]] - Merge similar configurations
• [[security-pattern-detection]] - Identify security anti-patterns
• [[compliance-checking]] - Ensure extracted patterns meet compliance requirements

Advanced Techniques

Multi-Stage Extraction Pipeline

# Stage 1: Initial scan
extract patterns from ./infra/ --stage scan

# Stage 2: Pattern classification
extract patterns from ./infra/ --stage classify

# Stage 3: Variable inference
extract patterns from ./infra/ --stage variables

# Stage 4: Template generation
extract patterns from ./infra/ --stage generate

# Stage 5: Validation
extract patterns from ./infra/ --stage validate

Machine Learning-Enhanced Extraction

Use ML models to improve pattern detection:

• Clustering: Group similar configurations
• Anomaly Detection: Identify outliers for manual review
• Type Inference: Predict variable types from usage
• Dependency Analysis: Extract implicit dependencies

Cross-Repository Pattern Mining

Extract patterns across multiple repositories:

extract patterns --repos-file ./repos.txt \
  --output ./org-templates/ \
  --consolidate-org-wide

Output Formats

JSON Schema

{
  "template": "aws-ec2-instance",
  "version": "1.0.0",
  "variables": {
    "instance_type": {
      "type": "string",
      "default": "t3.medium",
      "description": "EC2 instance type"
    }
  }
}

YAML Schema

template: aws-ec2-instance
version: 1.0.0
variables:
  instance_type:
    type: string
    default: t3.medium
    description: EC2 instance type

Terraform Variable Definitions

variable "instance_type" {
  type        = string
  default     = "t3.medium"
  description = "EC2 instance type"
}

Integration Points

• CI/CD Pipelines: Auto-extract patterns on code changes
• Template Registries: Publish extracted templates
• Documentation Systems: Generate docs from patterns
• Monitoring: Track pattern usage and evolution
• Compliance Tools: Validate against org standards

Troubleshooting

Pattern Detection Issues

Problem: Too many false positives Solution: Increase confidence threshold, add exclusion patterns

Problem: Missing obvious patterns Solution: Check file encoding, adjust regex patterns, enable debug mode

Problem: Inconsistent variable names Solution: Enable smart naming normalization, use naming dictionary

Template Generation Issues

Problem: Templates too generic Solution: Use narrower extraction scope, increase specificity

Problem: Templates too specific Solution: Widen extraction scope, enable cross-file consolidation

Version: 1.0.0 Last Updated: 2026-01-19 Skill Type: Analysis & Extraction Complexity: Advanced