Guide · Agent skill
How to use the Chief AI Officer Advisor agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Chief AI Officer Advisor is an agent skill for startup founders and CTOs who need a structured first pass on AI strategy. It helps the agent compare calling a model API, fine-tuning a smaller model, or building in-house, screen AI use cases for regulatory risk, estimate when self-hosting becomes cheaper than paying per token, and suggest the next AI role to hire. After this guide you will know how to install the skill, trigger it with plain requests and run its three included Python scripts.
Chief AI Officer Advisoralirezarezvani/claude-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the skill ZIP from this page.
- 2Install the skill. Unzip it into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder for one project. In the Claude apps, upload it under Settings → Capabilities → Skills.
- 3Describe your decision. Ask in plain language, naming the use case, expected volume and region, such as whether to call an API or fine-tune.
- 4Provide a JSON file if you have real numbers. Save your use case or workload as a JSON file so the scripts use your figures instead of the embedded sample.
- 5Run the matching script. Run scripts/model_buildvsbuy_calculator.py, scripts/ai_risk_classifier.py or scripts/ai_cost_economics.py with your file path.
- 6Check the output. Verify cost assumptions against current prices and have qualified counsel review any regulatory classification.
What an agent skill is and what this one makes the agent do
An agent skill is a folder of instructions and reference files that an AI agent loads when a task matches the skill's description. This one, named chief-ai-officer-advisor, gives the agent a strategic advisor role for AI decisions rather than a coding role.
It covers four decisions: whether to use an API, fine-tune or build your own model (with a 3-year total cost of ownership comparison); how an AI use case classifies under the EU AI Act and US state rules; when switching from API to self-hosted inference pays off; and which AI role to hire next. It is explicitly strategic and points engineering work such as RAG, prompt engineering, agent design and model deployment to other skills.
- Model build-vs-buy: API vs fine-tune vs in-house, with a use-case-specific recommendation
- AI risk classification under EU AI Act, NIST AI RMF and US state laws, with a required-controls list
- AI cost economics: API-to-self-hosted breakeven analysis
- AI team org evolution: stage-to-role hiring guidance
How to install the Chief AI Officer Advisor skill
Download the ZIP for this skill from this page. You can install it in one of two ways, depending on how you use your agent.
For Claude Code, unzip the download into ~/.claude/skills/ to make it available across your projects, or into a project's .claude/skills/ folder to limit it to that project. For the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
- Personal install: unzip into ~/.claude/skills/
- Single-project install: unzip into the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
~/.claude/skills/How to trigger the skill with natural requests
You do not need a special command. The skill loads when your request matches its description. Describe the decision in plain language and mention the topic. The skill's description names triggers such as CAIO, AI strategy, model selection, foundation model, fine-tuning, EU AI Act, NIST AI RMF, AI governance, model risk and AI economics.
Give concrete context in the request: the use case, the expected monthly volume, the country where the product will be used and the stage of the company. The more specific the input, the more useful the first-pass answer.
- Should we call an API or fine-tune a model for our support tagging feature?
- Is our resume-screening feature high-risk under the EU AI Act?
- At roughly 50M tokens a day, when does self-hosting an open model beat our API bill?
- Which AI role should we hire next at Series A, an AI engineer or an ML research scientist?
Run the included scripts for a numeric answer
The skill ships three Python scripts. Each one runs on an embedded sample when called with no arguments. To analyse your own case, pass the path to a JSON file describing your use case or workload. The scripts use only the Python 3 standard library.
The build-vs-buy calculator produces a use-case-specific recommendation with 3-year TCO. The risk classifier produces a risk tier and required-controls list. The cost economics script runs the API vs self-hosted comparison. Treat the outputs as a starting point, not a forecast.
- Build-vs-buy: model_buildvsbuy_calculator.py, optional use_case.json
- Risk screen: ai_risk_classifier.py, optional use_case.json
- Breakeven: ai_cost_economics.py, optional workload.json
python scripts/model_buildvsbuy_calculator.py path/to/use_case.json
python scripts/ai_risk_classifier.py path/to/use_case.json
python scripts/ai_cost_economics.py path/to/workload.jsonWhat the skill needs and its safety notes
Requirements are modest. You need Python 3 for the scripts and, optionally, JSON input files describing your use case or workload. The skill is documented as not using the network, and its scripts run locally.
The scripts' built-in assumptions and sample data may not match real GPU or vendor prices, so do not treat cost outputs as precise forecasts. The risk classifier is described as not legal advice. Have qualified AI counsel review any regulatory conclusion before you launch in the EU or US.
- Requires Python 3; scripts use only the standard library
- Runs scripts: yes. Uses network: no
- Cost figures use built-in assumptions; check them against current vendor and GPU prices
- Regulatory output is a screen, not legal advice
Tips for getting useful results
Start with the questions the skill asks back: what the AI must be good at, how you will measure it, and what happens when the model is wrong. Without an eval set and an error tolerance, the recommendation has little to anchor on.
Replace the sample data with your own numbers before trusting the economics. Use the reference files in the skill folder to read the decision logic behind a recommendation.
- Describe your measurement method and error tolerance up front
- Pass your own JSON file instead of the embedded sample
- Read the reference files for the full decision tree and regulatory landscape
- Ask for engineering help separately, since RAG and deployment are out of scope
What people use it for
- A startup founder without a Chief AI Officer comparing an API, a fine-tuned smaller model and an in-house build for a customer-support feature.
- A product lead screening a hiring or credit-related AI feature for likely high-risk obligations under the EU AI Act and US state rules before launch.
- A CTO estimating whether monthly token volume is high enough to justify moving an open model to self-hosted GPUs.
- A founder deciding whether the next AI hire should be an AI engineer or an ML research scientist, given the product's current stage.
Frequently asked questions
Is the Chief AI Officer Advisor skill free?
The skill is published under the MIT licence, which permits use, modification and redistribution subject to the licence terms. Check the licence file in the download for the full terms.
How do I install the Chief AI Officer Advisor skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for all projects, or into a project's .claude/skills/ folder for one project. In the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
What do I need to run the scripts?
You need Python 3. The scripts use only the Python standard library, and you can optionally pass a JSON file describing your use case or workload.
Does the skill send my data over the network?
The skill documentation states that it does not use the network. Its scripts are documented as running locally, and you can pass your own JSON input files.
Can the skill give me legal advice on the EU AI Act?
No. The risk classifier is described as not legal advice. Use its output as a first screen and have qualified AI counsel review the classification before launch.
Does this skill help with RAG, prompts or deployment?
No. The skill is strategic only and does not cover tactical engineering such as RAG, agent design, prompt engineering, evals or model deployment. It points those tasks to other skills.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits founders or CTOs without a Chief AI Officer who want a structured first pass on model strategy, a rough regulatory risk screen or a breakeven estimate for self-hosting. Do not use it for legal compliance decisions, hands-on engineering work or precise cost forecasts, and pick a qualified advisor or an engineering skill for those.
Download the skill