Guide · Agent skill
How to use the CTO Architecture Review agent skill (cto-review)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The cto-review skill is a set of instructions that makes an AI agent (Claude) act like a demanding CTO when you bring it an architecture, platform choice or scaling plan. It asks six hard questions covering where the system breaks under load, tech debt, team ramp, build versus buy, reliability targets and security exposure. This guide explains what an agent skill is, how to install this one, how to trigger it with normal requests, and what to watch for before relying on its output.
CTO Architecture Reviewalirezarezvani/claude-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the cto-review ZIP from this page and unzip it. The folder contains SKILL.md.
- 2Place the skill folder. Move it to ~/.claude/skills/ for all projects, or to .claude/skills/ inside one project.
- 3Or upload in the Claude apps. Go to Settings → Capabilities → Skills and upload the ZIP.
- 4Check Python and plugin files. Confirm Python is installed and that the claude-skills plugin folders, including vendor/c-level-advisor, are present.
- 5Invoke the skill. Run /c-level-agents:cto-review with your plan, or describe the architecture decision in plain language.
- 6Provide your data. Supply current load, growth, SLO and vendor details so the questions can be answered with specifics.
- 7Apply the verdict. Act on the SHIP, SHARPEN or BLOCK verdict and the three Next Steps, and route financial or legal questions to their own reviews.
What an agent skill is and what cto-review makes the agent do
An agent skill is a folder containing a SKILL.md file. The file tells the agent when to use the skill and how to structure its work. When a request matches the skill's description, the agent follows its instructions instead of answering generically.
Once cto-review is installed, the agent reviews a proposed architecture, platform choice or scaling plan by asking six questions: where the design breaks at 10x load, the top tech debt item and its cost, team ramp time, whether building beats buying, the reliability (SLO) targets, and what security exposure it creates. It then fills in a structured review ending in a SHIP, SHARPEN or BLOCK verdict.
- Scaling cliff: where the current architecture breaks in terms of users, requests or data volume
- Tech debt inventory: top item, weekly cost and when it becomes blocking
- Team scaling: ramp time and contribution model for open roles
- Build vs buy: why build, and the three-year cost of each option
- SLO and reliability: defined SLOs and error budget burn
- Security and compliance surface, with a sign-off check
Install the cto-review skill
Download the ZIP from this page and unzip it. To make the skill available across all your projects, place the folder in ~/.claude/skills/. To limit it to a single project, place it in that project's .claude/skills/ folder.
If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the ZIP there.
- Personal (all projects): ~/.claude/skills/
- Single project: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
Make sure the required files and Python are present
The skill's SKILL.md references two bundled Python scripts, and it needs Python to run them. The referenced scripts and links only resolve if the surrounding claude-skills plugin folders (vendor/c-level-advisor and related skills) are also present. If you install the single folder alone, expect the scripts and cross-references to fail.
The skill's own script paths use the plugin root variable, so they point into the plugin's vendor/c-level-advisor/cto-advisor/scripts/ folder.
- Python installed, to run tech_debt_analyzer.py and team_scaling_calculator.py
- The full claude-skills plugin folder structure, including vendor/c-level-advisor
- The skill does not use the network, according to its listing
Trigger the skill with a natural request
You can invoke the skill with the slash command /c-level-agents:cto-review followed by your plan. You can also describe the situation in plain language. The skill's description says to use it when committing to an architecture, planning for 10x load, or weighing a rebuild against a vendor, so requests like those should match it.
Give the agent the details it needs: the current system, expected growth, the database and services involved, and any vendor you are considering. Vague input produces vague answers, because the skill asks for specific metrics.
- Name the plan explicitly: what you want to build or switch to
- Supply current load, growth rate and error or SLO data if you have them
- State the decision you need: approve, revise or stop
/c-level-agents:cto-review <plan>Run the scripts and read the structured review
The workflow has four steps: run the tech debt analyzer and team scaling calculator, define the scaling-cliff hypothesis explicitly, cross-check security implications with cs-ciso-advisor, then apply the verdict.
The output is a markdown review with sections for Scaling Cliff, Tech Debt, Team, Build vs Buy, Reliability, Security, Verdict and Next Steps. The scripts' numbers depend on the inputs you provide, so treat their output as a starting point for your own data rather than a fixed forecast.
- Verdict options: 🟢 SHIP, 🟡 SHARPEN, 🔴 BLOCK
- Build vs buy outputs BUILD or BUY
- Next Steps lists three concrete actions
python ${CLAUDE_PLUGIN_ROOT}/vendor/c-level-advisor/cto-advisor/scripts/tech_debt_analyzer.py
python ${CLAUDE_PLUGIN_ROOT}/vendor/c-level-advisor/cto-advisor/scripts/team_scaling_calculator.pySafety notes on scripts and scope
This skill runs scripts, so read them before running them on a machine with sensitive data. The listing states it runs scripts but does not use the network. Confirm the script behaviour yourself if that matters to your organisation.
The skill does not replace financial or legal review. For build-vs-buy decisions above $100K a year, the skill routes you to cfo-review, and for data-surface changes it points to ciso-review.
- Runs scripts: yes. Review the two Python files before use.
- Uses network: no, per the listing
- Financial and legal decisions still need their own review
Tips for getting a useful review
Bring numbers. The skill's own guidance is to be specific, and to run a load test if you do not know where the system breaks. Without an SLO, it says, you cannot reason about reliability tradeoffs, so define one first or ask for help with the slo-architect skill.
Use it at decision points, not for every change. It is designed for committing to architecture, planning for 10x load, and weighing a rebuild against a vendor.
- Write the plan as a short document before asking for the review
- Include current capacity, growth and any missed SLOs
- Ask for the three Next Steps to be turned into tickets
What people use it for
- Deciding whether to approve a database or language change before a major rollout
- Pressure-testing a plan to handle ten times current traffic
- Comparing a rebuild of an internal system against buying a vendor product
- Checking the security and compliance surface of a new architecture before commit
- Planning whether doubling the engineering team will ramp in time
Frequently asked questions
Is the cto-review skill free?
The skill is published under the MIT licence, as stated on this page. Check the repository's licence file for the full terms before you use or redistribute it.
What do I need to install cto-review?
You need the skill folder containing SKILL.md, placed in ~/.claude/skills/ or a project's .claude/skills/ folder, or uploaded through Settings → Capabilities → Skills. You also need Python for the bundled scripts and the surrounding claude-skills plugin folders for the references to resolve.
Where do I put the cto-review folder?
Put it in ~/.claude/skills/ to make it available in all projects, or in .claude/skills/ inside a single project. In the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
Does cto-review run code on my computer?
Yes. The skill can run two bundled Python scripts, tech_debt_analyzer.py and team_scaling_calculator.py. Its listing says it does not use the network, but you should read the scripts before running them on sensitive systems.
Why does the skill say the scripts are missing?
The referenced scripts and links depend on the surrounding claude-skills plugin folders, such as vendor/c-level-advisor and related skills. If only the cto-review folder is installed, those paths will not resolve, so install the full plugin structure.
Can cto-review replace a CFO or legal review?
No. The skill does not replace financial or legal review. It routes build-vs-buy decisions above $100K a year to cfo-review and data-surface changes to ciso-review.
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
cto-review suits engineering leads and founders who need a structured, hard-questioned review before approving an architecture change, platform switch or large hiring plan. It is less useful for tiny projects or prototypes, where six questions add overhead, and it will not give reliable cost or capacity figures unless you supply your own data.
Download the skill