Cost Optimizer analyzes your cloud infrastructure spending across AWS, Azure, and GCP, identifying optimization opportunities through right-sizing, reserved instance recommendations, unused resource cleanup, and architectural suggestions. It provides specific, actionable recommendations with estimated savings and implementation effort.
Cloud bills have a way of growing faster than revenue. Cost Optimizer provides the visibility and recommendations you need to keep cloud spending efficient without sacrificing performance or reliability.
The skill connects to your cloud accounts (read-only), analyzes resource utilization, billing data, and usage patterns, then generates specific recommendations with estimated savings. It distinguishes between quick wins (unused resources, oversized instances) and strategic optimizations (reserved instances, architectural changes).
Unlike simple cost dashboards, it uses AI to understand your workload patterns and make contextual recommendations. An instance that’s idle at night might be a candidate for scheduling, while one that’s consistently at 10% CPU needs right-sizing.
# Analyze AWS account
cost-optimize --provider aws --profile production
# Multi-cloud analysis
cost-optimize --providers aws,azure,gcp --output report.html
# Focus on quick wins
cost-optimize --provider aws --min-savings 100 --max-effort low
Cloud Cost Analysis: AWS Production Account
Period: March 2026 | Total Spend: $47,230
💰 Optimization Opportunities:
Quick Wins (implement today, save $4,200/mo):
1. 🗑️ 12 unattached EBS volumes — $380/mo
2. 📉 3 idle RDS instances (0 connections) — $890/mo
3. 📦 S3 lifecycle policies for 40TB old logs — $620/mo
4. 🔄 Right-size 8 EC2 instances (avg 12% CPU) — $2,310/mo
Strategic (1-2 weeks, save $8,400/mo):
5. 💳 Reserved Instances for 15 stable workloads — $6,200/mo
6. 🏗️ Spot instances for batch processing — $2,200/mo
Total Potential Savings: $12,600/mo (26.7% of spend)
Annual Savings: $151,200
AI agents that work well with Cost Optimizer.
Official Chrome DevTools MCP server — AI agents can debug, profile, inspect DOM, and analyze web performance.
GitHub's official MCP server — interact with repos, issues, PRs, code search, and notifications via AI agents.
Official AWS MCP servers — AI agents interact with S3, Lambda, EC2, CloudFormation, Bedrock, and more.