cloud-finops-skills

# Cloud FinOps Skill & MCP

> Open-source FinOps knowledge skill and MCP server for AI agents - Claude, ChatGPT,
> Gemini, Cursor, and any MCP-compatible client. Cloud cost optimisation across AWS,
> Azure, GCP and OCI, AI inference economics, Kubernetes, data platforms, allocation,
> chargeback, anomaly management, and named-pattern waste detection playbooks.
> Built by [OptimNow](https://optimnow.io), grounded in enterprise delivery experience.

[![GitHub Stars](https://img.shields.io/github/stars/OptimNow/cloud-finops-skills?style=flat)](https://github.com/OptimNow/cloud-finops-skills/stargazers)
[![PyPI](https://img.shields.io/pypi/v/cloud-finops-mcp?label=cloud-finops-mcp)](https://pypi.org/project/cloud-finops-mcp/)
[![Latest release](https://img.shields.io/github/v/release/OptimNow/cloud-finops-skills?label=release)](https://github.com/OptimNow/cloud-finops-skills/releases/latest)
[![FinOps Framework](https://img.shields.io/badge/FinOps-Framework-blue)](https://www.finops.org/framework/)
[![Agent Skills](https://img.shields.io/badge/Agent-Skills%20Spec-green)](https://agentskills.io/specification)
[![Kiro Power](https://img.shields.io/badge/Kiro-Power-orange)](https://kiro.dev/docs/powers/installation/)
[![License: CC BY-SA 4.0](https://img.shields.io/badge/License-CC%20BY--SA%204.0-lightgrey.svg)](https://creativecommons.org/licenses/by-sa/4.0/)

---

## Install in 5 seconds

| Tool | One-step install |
|---|---|
| <img src="https://img.shields.io/badge/-Claude%20Code-D97757?logo=anthropic&logoColor=white" alt="Claude Code" height="22"/> | At the Claude Code prompt: `/plugin marketplace add https://github.com/OptimNow/cloud-finops-skills.git` then `/plugin install cloud-finops@optimnow` |
| <img src="https://img.shields.io/badge/-Claude.ai%20%2F%20Desktop-D97757?logo=anthropic&logoColor=white" alt="Claude.ai / Claude Desktop" height="22"/> | [Download the latest release zip](https://github.com/OptimNow/cloud-finops-skills/releases/latest), then **Settings -> Skills -> Upload zip** |
| <img src="https://img.shields.io/badge/-ChatGPT-10A37F?logo=openai&logoColor=white" alt="ChatGPT" height="22"/> | Self-host: `./install.sh --tool chatgpt --grouped` _(a public Cloud FinOps GPT is on the Roadmap)_ |
| <img src="https://img.shields.io/badge/-Gemini-4285F4?logo=googlegemini&logoColor=white" alt="Gemini" height="22"/> | Self-host: `./install.sh --tool gemini` _(a public Cloud FinOps Gem is on the Roadmap)_ |
| <img src="https://img.shields.io/badge/-Cursor-000000?logo=cursor&logoColor=white" alt="Cursor" height="22"/> <img src="https://img.shields.io/badge/-Windsurf-3DDC91?logoColor=white" alt="Windsurf" height="22"/> <img src="https://img.shields.io/badge/-Codex-412991?logo=openai&logoColor=white" alt="Codex" height="22"/> <img src="https://img.shields.io/badge/-Aider-0F172A?logoColor=white" alt="Aider" height="22"/> <img src="https://img.shields.io/badge/-Copilot-181717?logo=githubcopilot&logoColor=white" alt="Copilot" height="22"/> <img src="https://img.shields.io/badge/-Kiro%20IDE-FF6F00?logoColor=white" alt="Kiro IDE" height="22"/> <img src="https://img.shields.io/badge/-Gemini%20CLI-4285F4?logo=googlegemini&logoColor=white" alt="Gemini CLI" height="22"/> | One-liner: `curl -sL https://raw.githubusercontent.com/OptimNow/cloud-finops-skills/main/install.sh \| bash -s -- --tool <name>` |
| <img src="https://img.shields.io/badge/-Auto--detect-555555?logo=gnubash&logoColor=white" alt="Auto-detect" height="22"/> | `curl -sL https://raw.githubusercontent.com/OptimNow/cloud-finops-skills/main/install.sh \| bash` |
| <img src="https://img.shields.io/badge/-MCP%20hosted-7C3AED?logoColor=white" alt="MCP hosted" height="22"/> | Nothing to install: `claude mcp add --transport http cloud-finops https://cloud-finops-skills-590a051d.alpic.live/mcp`. For Claude.ai / Desktop, **Settings -> Connectors -> Add custom connector** with the same URL |
| <img src="https://img.shields.io/badge/-MCP%20package-7C3AED?logoColor=white" alt="MCP package" height="22"/> | `pip install cloud-finops-mcp` then add to your MCP client config (Claude Code / Cursor / Codex / Windsurf / Cline). Snippets: `./install.sh --tool mcp`. Six tools - faceted retrieval over the reference library and named-pattern playbooks. |

Full options, troubleshooting, and the model-agnostic API loader: see [INSTALLATION.md](./INSTALLATION.md).

### Live prices come from OptimToken, not from this repo

This skill carries billing **mechanics**, which stay true for years. It deliberately does
not carry current price **figures**, which go stale inside a packaged skill within weeks.

Those live in **[OptimToken](https://optimtoken.optimnow.io)** - LLM token rates for 250+
models and compute instance rates across seven clouds, each figure carrying its own as-of
date. The website works on its own with no setup. Adding its MCP connector lets the model
fetch a rate mid-answer instead of telling you to go and look it up:

| | |
|---|---|
| <img src="https://img.shields.io/badge/-OptimToken%20web-7C3AED?logoColor=white" alt="OptimToken web" height="22"/> | [optimtoken.optimnow.io](https://optimtoken.optimnow.io) - compare model and instance prices in the browser |
| <img src="https://img.shields.io/badge/-OptimToken%20MCP-7C3AED?logoColor=white" alt="OptimToken MCP" height="22"/> | Hosted, nothing to install. Point your client at `https://ai-pricing-hub-mcp-9604f763.alpic.live/` - config snippets in [INSTALLATION.md](./INSTALLATION.md#companion-connector-optimnow-ai-pricing-hub-optional) |

Pair it with the skill and a pricing question gets answered with a dated figure and its
source, rather than from a number the model remembers.

---

## What is a Skill, and why does it matter

A Skill is a structured knowledge file that you attach to an AI agent or a large language
model. It gives the model accurate, domain-specific context that it would not otherwise
have access to.

Without it, general-purpose LLMs make confident but incorrect statements on FinOps topics.
They miscalculate PTU break-even rates. They confuse Azure and AWS reservation mechanics.
They give generic advice that ignores how billing actually works on Bedrock or Azure OpenAI.
The answers sound plausible. Most of the time, they are wrong on the details that matter.

This skill corrects that by injecting verified, curated FinOps knowledge directly into the
model's context - covering billing models, cost allocation patterns, optimisation
frameworks, and governance practices across the major cloud providers and AI platforms.

**The closest analogy is RAG (Retrieval-Augmented Generation).** Like RAG, it extends a
model's knowledge beyond its training data. Unlike RAG, it requires no vector database,
no embedding pipeline, and no retrieval infrastructure. You copy a folder into your agent
setup and the model gains structured expertise on cloud financial management.

This makes it portable: the same skill works with Claude, GPT, Gemini, or any
MCP-compatible agent - with no changes to the files.

To keep responses consistent across models, add a **response contract** to your
system prompt (see `INSTALLATION.md`, "API integration / Recommended response
contract"). This ensures structured, billing-grounded answers even when model
defaults differ.

---

## Who this is for

- **FinOps practitioners** building or evaluating AI-assisted cost analysis tools
- **Cloud engineers and architects** who want a cost-aware assistant integrated into
  their workflow
- **Developers** building internal FinOps agents, chatbots, or automation pipelines
- **Finance and IT managers** evaluating the AI tooling their teams are deploying

No AI infrastructure experience is required to use this skill. If you can copy a folder
and follow the installation steps, you can add FinOps expertise to any compatible agent.

---

## Design principles

- **AI cost management is a first-class domain.** Most FinOps resources treat AI
  workloads as an edge case. This skill treats them as a primary concern, with
  dedicated reference files for each major AI platform.
- **Visibility before optimisation.** The skill follows a consistent sequence:
  establish what you are spending, understand what is driving it, then act. It does
  not recommend optimisation steps before the visibility preconditions are met.
- **Provider-mechanics-first, vendor-claim-skeptical.** Guidance is grounded in how
  billing actually works (CUR columns, Azure cost-management semantics, BigQuery
  export, FOCUS conformance) rather than in vendor marketing or framework
  positioning. Vendor sustainability and savings claims are read critically, with
  primary sources cited.
- **Maturity is contextual, not aspirational.** Verticals where cloud is not a
  revenue generator do not need to reach Run; Crawl plus selective Walk is the right
  state when cloud is a cost centre. Verticals where cloud IS the product need Run
  because cloud efficiency directly drives gross margin. Pushing every organisation
  toward the same maturity ceiling is malpractice.
- **Connect cost to business value.** Every recommendation answers the CFO test:
  what business outcome does this protect or unlock. Cost reduction without a value
  lens is a leak.
- **Mechanics live here, price figures do not.** Billing mechanics are durable and are
  what the reference files carry. Absolute prices are volatile, and a figure frozen in a
  markdown file goes stale within weeks with nothing in the distribution chain to correct
  it. So the skill routes current-price questions to a live source - the
  [OptimNow AI Pricing Hub](https://optimtoken.optimnow.io), which serves LLM token rates
  and compute instance rates across seven providers, each with its own as-of date - and
  states the date and source on any figure it does quote.
- **FinOps is an operating discipline, not a culture.** The discipline lives in
  allocation, anomaly management, commitment management, rightsizing, and
  governance, all of which produce measurab