finops-expert — скилл для Claude Code: анализ облачных затрат
Скилл finops-expert для Claude Code разбирает расходы на облако и ищет варианты оптимизации. Анализирует затраты AWS, прогнозирует расходы и даёт рекомендации по EC2.
Apache-2.0 239 установок ★ 45 обновлён 2026-10-12
npx skills add personamanagmentlayer/pcl --skill finops-expert -a claude-codeЧто делает finops-expert
finops-expert охватывает видимость облачных затрат, распределение расходов, бюджеты и прогнозирование. Он также разбирает оптимизацию ресурсов: Reserved Instances, Savings Plans, Spot instances, изменение размеров ресурсов и очистку простаивающих ресурсов. Это рекомендации и примеры для FinOps-задач, а не автоматическое применение изменений в облаке.
В примере для AWS класс AWSCostAnalyzer обращается к Cost Explorer: группирует затраты по сервисам, строит прогноз и получает рекомендации по размеру EC2. CostOptimizer ищет остановленные EC2-инстансы, незанятые EBS-тома и Elastic IP без привязки к инстансу. В коде используются boto3 и pandas; для запросов нужны доступ к AWS и разрешения на соответствующие API. Расчёт экономии Reserved Instances помечен как упрощённый и не использует фактический API цен — его оценку нельзя считать точной сметой.
Какие задачи решает
- разобрать расходы AWS по сервисам
- спрогнозировать затраты на облако
- найти простаивающие EC2-инстансы и незанятые EBS-тома
- сравнить расходы On-Demand и Reserved Instances по упрощённой модели
Примеры запросов агенту
- Разбери расходы AWS по сервисам и покажи самые затратные.
- Найди незанятые EC2-инстансы, EBS-тома и Elastic IP.
- Спрогнозируй расходы AWS и объясни рекомендации по изменению размера EC2.
Как установить finops-expert
1Через Skills CLI (Vercel)
npx skills add personamanagmentlayer/pcl --skill finops-expert -a claude-code2Вручную: положить файлы
~/.claude/skills/finops-expert/SKILL.mdПапка скилла целиком; для одного проекта — .claude/skills/finops-expert/
1Через Skills CLI (Vercel)
npx skills add personamanagmentlayer/pcl --skill finops-expert -a codex2Вручную: положить файлы
.agents/skills/finops-expert/SKILL.mdДля всех проектов — ~/.agents/skills/
1Через Skills CLI (Vercel)
npx skills add personamanagmentlayer/pcl --skill finops-expert -a cursor2Вручную: положить файлы
.cursor/skills/finops-expert/SKILL.mdCursor читает и .agents/skills/, и .claude/skills/
1GitHub CLI
gh skill install personamanagmentlayer/pcl stdlib/professional/finops-expert2Вручную: положить файлы
.github/skills/finops-expert/SKILL.md1Через Skills CLI (Vercel)
npx skills add personamanagmentlayer/pcl --skill finops-expert -a gemini-cli2Командой платформы
gemini skills install https://github.com/personamanagmentlayer/pcl --path stdlib/professional/finops-expertВручную: положить файлы
.devin/skills/finops-expert/SKILL.mdСтарый путь .windsurf/skills/ тоже читается
Через Skills CLI (Vercel)
npx skills add personamanagmentlayer/pcl --skill finops-expert -a opencodeClaude.ai
Упакуйте папку в zip и загрузите в Settings → Capabilities → Skills
По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов
Вопросы о finops-expert
Как установить finops-expert в Claude Code?
Выполните npx skills add personamanagmentlayer/pcl --skill finops-expert -a claude-code. Скилл размещается в ~/.claude/skills/finops-expert/SKILL.md.
Что делает finops-expert?
Помогает анализировать облачные расходы, прогнозировать затраты и находить варианты оптимизации. Для AWS в примерах есть запросы к Cost Explorer и рекомендации по размеру EC2.
Что нужно для анализа расходов AWS?
Пример использует boto3 и обращается к AWS Cost Explorer, поэтому нужны доступ к AWS и разрешения для соответствующих API.
Исходный текст
SKILL.md · Apache-2.0
---
name: finops-expert
version: 2.0.0
description: >-
Expert-level cloud financial operations, cost optimization, and cloud economics. Use when
the user mentions cloud cost, optimization, cloud economics, or AWS cost, or when the
task involves FinOps Fundamentals, Cost Management, FinOps Practices, or Cost Visibility.
category: professional
tags: [finops, cloud-cost, optimization, cloud-economics, aws-cost]
allowed-tools:
- Read
- Write
- Edit
- Bash(python:*, python3:*, pip:*, pytest:*)
---
# FinOps Expert
Expert guidance for cloud financial operations, cost optimization, resource management, and cloud economics.
## Core Concepts
### FinOps Fundamentals
- Cloud cost visibility
- Usage optimization
- Rate optimization
- Architecture optimization
- Cloud unit economics
- Showback and chargeback
### Cost Management
- Reserved Instances (RIs)
- Savings Plans
- Spot instances
- Right-sizing resources
- Idle resource cleanup
- Storage lifecycle policies
### FinOps Practices
- Tagging strategies
- Budgets and alerts
- Cost allocation
- Forecasting and planning
- Cross-team collaboration
- Continuous optimization
## AWS Cost Analysis
```python
import boto3
from datetime import datetime, timedelta
from typing import Dict, List
import pandas as pd
class AWSCostAnalyzer:
"""Analyze AWS costs using Cost Explorer API"""
def __init__(self):
self.ce_client = boto3.client('ce')
def get_cost_and_usage(self, start_date: str, end_date: str,
granularity: str = 'DAILY',
metrics: List[str] = None) -> Dict:
"""Get cost and usage data"""
if metrics is None:
metrics = ['UnblendedCost', 'UsageQuantity']
response = self.ce_client.get_cost_and_usage(
TimePeriod={
'Start': start_date,
'End': end_date
},
Granularity=granularity,
Metrics=metrics,
GroupBy=[
{'Type': 'DIMENSION', 'Key': 'SERVICE'}
]
)
return response['ResultsByTime']
def get_top_services_by_cost(self, days: int = 30, top_n: int = 10) -> pd.DataFrame:
"""Get top services by cost"""
end_date = datetime.now().strftime('%Y-%m-%d')
start_date = (datetime.now() - timedelta(days=days)).strftime('%Y-%m-%d')
results = self.get_cost_and_usage(start_date, end_date, 'MONTHLY')
service_costs = {}
for result in results:
for group in result['Groups']:
service = group['Keys'][0]
cost = float(group['Metrics']['UnblendedCost']['Amount'])
if service in service_costs:
service_costs[service] += cost
else:
service_costs[service] = cost
df = pd.DataFrame(list(service_costs.items()),
columns=['Service', 'Cost'])
return df.nlargest(top_n, 'Cost')
def get_cost_forecast(self, days_ahead: int = 30) -> Dict:
"""Get cost forecast"""
start_date = datetime.now().strftime('%Y-%m-%d')
end_date = (datetime.now() + timedelta(days=days_ahead)).strftime('%Y-%m-%d')
response = self.ce_client.get_cost_forecast(
TimePeriod={
'Start': start_date,
'End': end_date
},
Metric='UNBLENDED_COST',
Granularity='MONTHLY'
)
return {
'forecasted_cost': float(response['Total']['Amount']),
'mean_value': float(response['ForecastResultsByTime'][0]['MeanValue'])
}
def get_rightsizing_recommendations(self) -> List[Dict]:
"""Get EC2 rightsizing recommendations"""
response = self.ce_client.get_rightsizing_recommendation(
Service='AmazonEC2'
)
recommendations = []
for rec in response['RightsizingRecommendations']:
recommendations.append({
'instance_id': rec['CurrentInstance']['ResourceId'],
'current_type': rec['CurrentInstance']['InstanceType'],
'recommended_type': rec['ModifyRecommendationDetail']['TargetInstances'][0]['InstanceType']
if rec.get('ModifyRecommendationDetail') else None,
'estimated_savings': float(rec['EstimatedMonthlySavings']['Value'])
if rec.get('EstimatedMonthlySavings') else 0
})
return recommendations
class CostOptimizer:
"""Optimize cloud costs"""
def __init__(self):
self.ec2_client = boto3.client('ec2')
self.rds_client = boto3.client('rds')
self.s3_client = boto3.client('s3')
def find_idle_resources(self) -> Dict[str, List]:
"""Find idle/unused resources"""
idle_resources = {
'ec2_instances': [],
'ebs_volumes': [],
'elastic_ips': [],
'load_balancers': []
}
# Idle EC2 instances (stopped for > 7 days)
instances = self.ec2_client.describe_instances(
Filters=[{'Name': 'instance-state-name', 'Values': ['stopped']}]
)
for reservation in instances['Reservations']:
for instance in reservation['Instances']:
idle_resources['ec2_instances'].append({
'id': instance['InstanceId'],
'type': instance['InstanceType'],
'state': instance['State']['Name']
})
# Unattached EBS volumes
volumes = self.ec2_client.describe_volumes(
Filters=[{'Name': 'status', 'Values': ['available']}]
)
for volume in volumes['Volumes']:
idle_resources['ebs_volumes'].append({
'id': volume['VolumeId'],
'size': volume['Size'],
'type': volume['VolumeType']
})
# Unattached Elastic IPs
addresses = self.ec2_client.describe_addresses()
for address in addresses['Addresses']:
if 'InstanceId' not in address:
idle_resources['elastic_ips'].append({
'allocation_id': address['AllocationId'],
'public_ip': address['PublicIp']
})
return idle_resources
def calculate_reserved_instance_savings(self,
instance_type: str,
count: int,
term: int = 1) -> Dict:
"""Calculate RI savings"""
# Simplified calculation (would use actual pricing API)
on_demand_hourly = self._get_on_demand_price(instance_type)
ri_hourly = on_demand_hourly * 0.65 # ~35% discount
hours_per_year = 24 * 365
annual_on_demand = on_demand_hourly * hours_per_year * count
annual_ri = ri_hourly * hours_per_year * count
return {
'instance_type': instance_type,
'count': count,
'annual_on_demand_cost': annual_on_demand,
'annual_ri_cost': annual_ri,
'annual_savings': annual_on_demand - annual_ri,
'savings_percentage': ((annual_on_demand - annual_ri) / annual_on_demand) * 100
}
def _get_on_demand_price(self, instance_type: str) -> float:
"""Get on-demand hourly price (simplified)"""
# In production, use AWS Pricing API
prices = {
't3.micro': 0.0104,
't3.small': 0.0208,
't3.medium': 0.0416,
'm5.large': 0.096,
'm5.xlarge': 0.192
}
return prices.get(instance_type, 0.10)
```
## Cost Allocation and Tagging
```python
class CostAllocation:
"""Manage cost allocation with tags"""
def __init__(self):
self.ec2_client = boto3.client('ec2')
self.ce_client = boto3.client('ce')
def define_tagging_strategy(self) -> Dict[str, List[str]]:
"""Define mandatory tags"""
return {
'environment': ['prod', 'staging', 'dev'],
'team': ['engineering', 'data', 'product'],
'cost_center': ['CC001', 'CC002', 'CC003'],
'project': ['project-a', 'project-b'],
'owner': ['email addresses']
}
def audit_resource_tags(self, resource_type: str = 'instance') -> List[Dict]:
"""Audit resources for missing tags"""
mandatory_tags = ['environment', 'team', 'cost_center']
untagged_resources = []
if resource_type == 'instance':
instances = self.ec2_client.describe_instances()
for reservation in instances['Reservations']:
for instance in reservation['Instances']:
tags = {tag['Key']: tag['Value']
for tag in instance.get('Tags', [])}
missing_tags = [tag for tag in mandatory_tags
if tag not in tags]
if missing_tags:
untagged_resources.append({
'resource_id': instance['InstanceId'],
'missing_tags': missing_tags
})
return untagged_resources
def get_cost_by_tag(self, tag_key: str, start_date: str,
end_date: str) -> pd.DataFrame:
"""Get costs grouped by tag"""
response = self.ce_client.get_cost_and_usage(
TimePeriod={
'Start': start_date,
'End': end_date
},
Granularity='MONTHLY',
Metrics=['UnblendedCost'],
GroupBy=[
{'Type': 'TAG', 'Key': tag_key}
]
)
costs = []
for result in response['ResultsByTime']:
for group in result['Groups']:
costs.append({
'tag_value': group['Keys'][0].split('$')[1]
if '$' in group['Keys'][0] else 'Untagged',
'cost': float(group['Metrics']['UnblendedCost']['Amount'])
})
return pd.DataFrame(costs)
```
## Budget Management
```python
class BudgetManager:
"""Manage AWS budgets and alerts"""
def __init__(self):
self.budgets_client = boto3.client('budgets')
self.account_id = boto3.client('sts').get_caller_identity()['Account']
def create_monthly_budget(self, name: str, amount: float,
email: str) -> Dict:
"""Create monthly cost budget with alerts"""
budget = {
'BudgetName': name,
'BudgetLimit': {
'Amount': str(amount),
'Unit': 'USD'
},
'TimeUnit': 'MONTHLY',
'BudgetType': 'COST'
}
# Alert at 80% and 100%
notifications = [
{
'Notification': {
'NotificationType': 'ACTUAL',
'ComparisonOperator': 'GREATER_THAN',
'Threshold': 80,
'ThresholdType': 'PERCENTAGE'
},
'Subscribers': [{
'SubscriptionType': 'EMAIL',
'Address': email
}]
},
{
'Notification': {
'NotificationType': 'ACTUAL',
'ComparisonOperator': 'GREATER_THAN',
'Threshold': 100,
'ThresholdType': 'PERCENTAGE'
},
'Subscribers': [{
'SubscriptionType': 'EMAIL',
'Address': email
}]
}
]
response = self.budgets_client.create_budget(
AccountId=self.account_id,
Budget=budget,
NotificationsWithSubscribers=notifications
)
return response
```
## Best Practices
### Cost Visibility
- Implement comprehensive tagging
- Enable Cost Explorer
- Set up cost allocation tags
- Create custom cost reports
- Use dashboards for visualization
- Monitor costs daily
### Optimization
- Right-size resources regularly
- Use Reserved Instances/Savings Plans
- Leverage Spot instances for flexible workloads
- Implement auto-scaling
- Clean up idle resources
- Use storage lifecycle policies
### Governance
- Set budgets and alerts
- Implement approval workflows
- Regular cost reviews
- Cross-team accountability
- Document cost optimization wins
- Automate cost controls
## Anti-Patterns
❌ No tagging strategy
❌ Ignoring rightsizing recommendations
❌ Not using Reserved Instances
❌ No budget alerts
❌ Keeping idle resources
❌ Manual cost tracking
❌ Siloed cost responsibility
## Resources
- AWS Cost Explorer: https://aws.amazon.com/aws-cost-management/aws-cost-explorer/
- FinOps Foundation: https://www.finops.org/
- AWS Well-Architected Cost Optimization: https://wa.aws.amazon.com/wat.pillar.costOptimization.en.html
- Boto3 Cost Explorer: https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/ce.html
© personamanagmentlayer · Apache-2.0 · personamanagmentlayer/pcl