CAIO Forcing Questions — Business

The CAIO Forcing Questions skill is an eval-demanding Chief AI Officer interrogation that pressure-tests any plan involving AI. It asks six critical questions covering eval discipline, hallucination SLOs, EU AI Act risk classification, build-vs-buy decisions, cost economics, and hiring prerequisites, producing a structured review with a verdict.

c-level-advisor claude-code

📂 View skill on GitHub ↗ 🚀 How to use

✨ What it does

Forces definition of eval criteria with 50-100 representative inputs and grading rubric.
Requires quantified hallucination/error SLOs and fallback mechanisms.
Classifies AI use case under EU AI Act tiers (PROHIBITED, HIGH, LIMITED, MINIMAL).
Runs model build-vs-buy calculator to recommend API, fine-tune, or build.
Projects 12-month cost trajectory and breakeven for self-hosted migration.
Maps AI capability to specific roles and checks prerequisite hires.

🎯 When to use it

Before shipping any new AI-powered feature without an eval set.
When choosing between API, fine-tune, or self-hosted model deployment.
Before EU launch of any AI feature to classify risk under the EU AI Act.
Before signing a multi-year AI vendor contract or committing to fine-tuning projects.
Before a major AI team hire, especially ML engineer or research scientist.

🚀 How to use

Trigger: /cs:caio-review

Trigger with the slash command /cs:caio-review <plan> where <plan> describes the AI initiative. The skill expects a plan that includes the AI use case, domain, and scale. It will guide you through six questions and optionally run scripts for model selection, risk classification, and cost economics. Example prompts:

/cs:caio-review We plan to add an AI chatbot for customer support in the EU.
/cs:caio-review We are considering fine-tuning a model for medical diagnosis.

📄 Output: A structured CAIO Review report with verdict (SHIP, SHARPEN, or BLOCK) and next steps.

📦 Add this skill to Claude Code

# 1. Get the skills repo
git clone --depth 1 https://github.com/alirezarezvani/claude-skills /tmp/claude-skills

# 2. Copy this skill into your project (or ~/.claude/skills for all projects)
mkdir -p .claude/skills
cp -r /tmp/claude-skills/c-level-agents/skills/caio-review .claude/skills/caio-review

Skill source: c-level-agents/skills/caio-review/SKILL.md

⚠️ Good to know

Requires access to the referenced scripts and the ability to run Python; the skill itself does not execute the analysis but guides the process.

❓ FAQ

What is the minimum eval set size required before shipping an AI feature?

The skill requires 50-100 representative inputs minimum, with expected outputs or a grading rubric.

What does the skill recommend for most B2B SaaS use cases?

It recommends using an API for 80% of B2B SaaS use cases, with fine-tuning for 15% and building from scratch for less than 1%.

When is a conformity assessment required under the EU AI Act?

A conformity assessment is required for HIGH-risk AI systems, which involves 10 Articles of obligations and can take 3-12 months and cost $50-200K.

🤖 Overview, features, install steps and FAQ were generated from the project's SKILL.md on Sep 4, 2026. Always check the original source before running commands.