cto-review

Overview of cto-review skill

What cto-review does

cto-review is a strategic engineering review skill for pressure-testing architecture, scaling, tech debt, hiring, and build-vs-buy decisions before they become expensive commitments. It is built around the command pattern /cs:cto-review <plan> and asks CTO-style forcing questions: where the system breaks, what debt is becoming blocking, whether the team can scale the plan, and whether buying is more rational than building.

Best fit for Strategic Planning decisions

The cto-review skill is most useful when you already have a concrete plan, not just a vague idea. Good use cases include approving a new database, planning for 10x traffic, deciding whether to rebuild a legacy system, evaluating a platform migration, or preparing a technical strategy review for leadership. It is especially relevant for Strategic Planning where engineering choices affect budget, hiring, reliability, and long-term execution risk.

What makes it different from a generic prompt

A generic AI prompt may produce broad architecture advice. cto-review focuses the conversation on six decision questions that expose scaling cliffs, weekly cost of tech debt, team ramp constraints, and three-year total cost of ownership. That structure makes it better for executive review, architecture approval, and pre-commitment risk analysis than open-ended brainstorming.

Important adoption notes

The repository path contains a single SKILL.md for this skill, with no local README.md, metadata.json, rules, resources, or bundled helper scripts in the skill folder. The source references analyzer scripts via relative paths under a broader cto-advisor area, so verify script availability in your installed environment before relying on those commands. The core value of cto-review is the review framework; automation may require extra setup.

How to Use cto-review skill

cto-review install and repository check

Install from the GitHub skill collection using your skill manager, for example:

npx skills add alirezarezvani/claude-skills --skill cto-review

After installing, open:

c-level-advisor/c-level-agents/skills/cto-review/SKILL.md

Read this file first because it contains the command, trigger conditions, and the six CTO questions. Also inspect the surrounding repository if your agent environment preserves relative paths, because the skill references scripts outside the local folder, such as cto-advisor/scripts/tech_debt_analyzer.py and cto-advisor/scripts/team_scaling_calculator.py.

Inputs the skill needs

For strong cto-review usage, provide a plan with enough operational detail for the skill to challenge assumptions. Include:

• Current scale: users, requests per second, data volume, tenants, regions, or transaction counts
• Target scale: expected growth, 10x scenario, launch event, enterprise expansion, or migration deadline
• Architecture: services, datastore choices, queues, dependencies, deployment model, and reliability targets
• Constraints: budget, team size, hiring plan, compliance, vendor limits, deadlines, and SLOs
• Decision under review: approve, delay, buy, build, rebuild, migrate, or run an experiment

Weak input: “Review our new architecture.”

Stronger input: “Use cto-review for Strategic Planning. We are moving from a monolith on PostgreSQL to event-driven services before expanding from 50k to 500k monthly active users. Current p95 API latency is 700ms during peak, writes are concentrated on one primary DB, the team has 8 engineers and plans to hire 4 more. Review scaling cliffs, tech debt cost, team ramp risk, and whether managed event streaming is better than building our own platform.”

Practical cto-review usage workflow

Start by pasting the plan after the command style used by the skill:

/cs:cto-review <your plan>

Ask for a decision-oriented output, not a general essay. Useful output formats include:

• “List the top 5 approval blockers.”
• “Identify the first scaling cliff and the evidence needed to confirm it.”
• “Compare build vs buy over three years.”
• “Separate reversible experiments from irreversible commitments.”
• “Give the questions I should ask the engineering lead before approval.”

If the first answer is too speculative, add missing facts rather than asking the model to “try again.” The skill is designed to interrogate a plan; it becomes more valuable as you supply real load data, cost estimates, staffing assumptions, and incident history.

Files and commands to verify first

The only confirmed skill file is SKILL.md. It includes example script calls, but this specific skill folder does not include local scripts. Before making the scripts part of your workflow, check whether these paths exist after installation:

../../../skills/cto-advisor/scripts/tech_debt_analyzer.py

../../../skills/cto-advisor/scripts/team_scaling_calculator.py

If they are missing, you can still use the cto-review guide manually by providing your own tech debt inventory and team ramp assumptions in the prompt.

cto-review skill FAQ

Is cto-review suitable for beginners?

Yes, if the user has a concrete engineering decision to review. It is not a beginner tutorial for system design, cloud architecture, or database selection. A junior founder or product leader can use it effectively by supplying business goals and asking the agent to translate technical risks into decision questions, but final judgment should involve an experienced engineer.

When should I not use cto-review?

Do not use cto-review for small implementation choices, routine code review, ticket grooming, or debugging a narrow defect. It is also a poor fit when no plan exists yet. Use it when the cost of being wrong is material: platform migration, high-cost vendor decision, scaling investment, reliability redesign, hiring plan, or rebuild proposal.

How does it compare with an architecture review prompt?

An architecture review prompt often checks design patterns, components, and tradeoffs. The cto-review skill adds executive pressure: cost per week of debt, team scalability, failure thresholds, and build-vs-buy economics. That makes it better for go/no-go review than for producing a full reference architecture from scratch.

Does cto-review require the referenced scripts?

No. The core cto-review skill is a structured review method in SKILL.md. The referenced scripts may improve analysis if available, but the skill can still run through prompts alone. Treat script support as optional until you confirm the files exist in your local installation.

How to Improve cto-review skill

Improve cto-review inputs before asking for judgment

The most common failure mode is asking for a CTO review without numbers. Add even rough ranges: “current write rate is 2k/min,” “expected peak is 10x after launch,” “vendor quote is $180k/year,” or “team has two backend engineers with no Kafka experience.” These details let cto-review identify real scaling cliffs instead of generic risks.

Ask for evidence gaps, not just recommendations

A strong cto-review guide should tell you what must be measured before approval. Ask: “What evidence is missing?” or “Which assumptions need a load test, cost model, prototype, or vendor reference call?” This turns the output into an action plan and reduces false confidence from plausible but unverified architecture advice.

Iterate with decision constraints

After the first review, narrow the decision. For example: “Assume we cannot hire until Q3,” “Assume downtime must stay under 30 minutes,” or “Assume build-vs-buy must be justified over three years.” Constraints force the skill to prioritize tradeoffs that matter to leadership rather than listing every possible technical concern.

Extend the skill for your organization

To improve cto-review in a team setting, add internal checklists for SLOs, incident history, cloud cost thresholds, approved vendors, security requirements, and architecture decision record templates. If you use the referenced analyzer scripts, document their paths and expected inputs near the skill so future users are not blocked by missing relative files.