cdo-review
by alirezarezvanicdo-review is a Chief Data Officer review skill for pressure-testing data strategy plans, AI training data rights, architecture choices, data productization, M&A diligence, and data-team hiring before commitments.
This skill scores 72/100, which means it is acceptable for directory listing but should be presented as a lightweight review checklist rather than a fully tooled workflow. Directory users can understand when to trigger it and what kind of CDO-style scrutiny it will apply, but they should expect limited implementation support beyond the SKILL.md guidance.
- Clear trigger and scope: `/cs:cdo-review <plan>` is explicitly for plans involving training data, data architecture, data productization, data hiring, or M&A diligence.
- Operationally useful framing: the skill supplies a Chief Data Officer pressure-test with six decision-oriented questions rather than a vague advisory prompt.
- Good install-decision clarity: the 'When to Run' section names specific scenarios such as ML training on customer data, data-infrastructure SaaS contracts, productizing customer data, and major data hires.
- No install command, README, references, scripts, or supporting materials are present, so adoption depends entirely on the single SKILL.md file.
- The provided evidence shows strong questioning prompts but little evidence of concrete output format, worked examples, or step-by-step execution beyond the review questions.
Overview of cdo-review skill
What cdo-review is for
cdo-review is a Chief Data Officer review skill for pressure-testing plans that depend on data strategy, data rights, data architecture, data monetization, AI training data, or data-team decisions. It is designed to be invoked as /cs:cdo-review <plan> and turns a broad proposal into a structured interrogation before the organization commits money, reputation, engineering time, or legal exposure.
Best-fit users and decisions
The cdo-review skill is most useful for founders, product leaders, data leaders, AI teams, and strategic planning groups evaluating plans such as training a model on customer data, buying a warehouse or lakehouse platform, launching a data product, hiring a senior data role, or reviewing data assets during M&A. It fits Strategic Planning work because it asks whether the data actually changes a business decision, not merely whether the idea sounds technically plausible.
What makes it different from a generic prompt
Instead of giving generic data-strategy advice, cdo-review uses forcing questions: what decision the data drives, what consent provenance exists, whether the proposed architecture matches the operating model, how data assets could be valued or misused, and what organizational capability is required. The value is not a polished memo; it is a sharper go/no-go review that exposes weak assumptions before they become contracts, roadmap commitments, or compliance problems.
Adoption cautions
The repository currently centers on SKILL.md only, with no extra scripts, references, rules, or metadata files to inspect. That makes cdo-review lightweight and easy to install, but users should not treat it as a full governance framework, legal review, privacy impact assessment, or vendor-selection model. It is best used as a decision review layer before deeper domain-specific diligence.
How to Use cdo-review skill
cdo-review install and repository path
Install from the skill repository with:
npx skills add alirezarezvani/claude-skills --skill cdo-review
The relevant source is under c-level-advisor/c-level-agents/skills/cdo-review/. Read SKILL.md first; there are no visible companion rules/, resources/, references/, or scripts/ folders in the current skill package. Because the skill is compact, installation decisions should focus on whether its review frame matches your data-strategy workflow.
What input the skill needs
For strong cdo-review usage, provide the plan as a decision package, not a slogan. Include the business decision being made, data sources, consent or contract basis, data classes, intended model or product use, expected users, architecture choices, vendor commitments, budget range, timeline, and who will own the result.
Weak prompt:
/cs:cdo-review We want to monetize customer data with AI.
Stronger prompt:
/cs:cdo-review Review a plan to create paid industry benchmarks from aggregated customer usage data. Sources: product telemetry and CRM records. Consent: current TOS mentions analytics but not resale. Buyers: enterprise customers. Architecture: Snowflake plus dbt. Timeline: 2 quarters. Decision needed: whether to approve product discovery and legal review before hiring a data PM.
Suggested cdo-review workflow
Use cdo-review before approval gates, not after implementation. A practical workflow is:
- Draft the plan in one page.
- Run
/cs:cdo-review <plan>. - Mark each challenge as
answered,unknown,risky, ornot applicable. - Rewrite the plan with missing provenance, decision logic, and ownership.
- Run the skill again against the revised version.
- Escalate unresolved consent, contractual, security, or valuation issues to the right specialist.
This makes the skill useful for strategic planning, vendor evaluation, AI training readiness, and data-product reviews without pretending it replaces legal, security, or finance diligence.
Prompt details that improve output quality
Name the decision you are trying to unblock. cdo-review is built around the question “What decision does this data drive?” so vague goals like “build a data moat” produce weaker output. Also separate first-party explicit opt-in, first-party TOS-only, third-party licensed, scraped, inferred, and customer-confidential data. The skill can then identify where provenance, consent, or intended use may break the plan.
cdo-review skill FAQ
Is cdo-review only for Chief Data Officers?
No. The cdo-review skill is written from a CDO perspective, but it is practical for founders, CEOs, CTOs, product managers, AI leads, data engineers, and investors. Anyone making a consequential data decision can use it to surface the questions a strong data executive would ask before approving the plan.
When should I not use cdo-review?
Do not use cdo-review as the final authority for privacy law, security architecture, model risk, tax, accounting, or M&A valuation. It is also a poor fit for purely operational tickets, small dashboard requests, or implementation debugging. Use it when the decision has strategic, legal, architectural, hiring, monetization, or trust implications.
How is cdo-review different from asking for a data strategy review?
A generic review may summarize pros and cons. cdo-review is more adversarial and decision-driven: it challenges whether the data should be collected, trained on, sold, licensed, centralized, decentralized, or staffed at all. That makes it better for pre-commitment review than for brainstorming broad possibilities.
Is cdo-review beginner-friendly?
Yes, if the user can describe the plan clearly. Beginners may need to gather basic facts first: where the data comes from, who consented, what systems store it, what decision depends on it, and what business outcome is expected. Without those facts, the skill will still help by showing what is missing.
How to Improve cdo-review skill
Improve cdo-review inputs before rerunning
The fastest way to improve cdo-review results is to add concrete evidence to the prompt. Include sample data-source lists, consent language summaries, vendor names, architecture diagrams in text form, ownership boundaries, retention assumptions, and revenue or cost expectations. The skill performs best when it can compare the stated business decision against actual data rights and operating constraints.
Watch for common failure modes
Common weak outputs come from weak plans: no named decision, unclear data provenance, mixed consent types, undefined customer benefit, architecture chosen before requirements, or hiring justified by title rather than capability gap. If the first review feels too abstract, rewrite the plan around a single approval question such as “Should we train this model?”, “Should we sign this vendor contract?”, or “Should we commercialize this dataset?”
Use iteration to turn critique into action
After the first cdo-review pass, ask for a decision table with columns for issue, risk, missing evidence, owner, and next action. Then revise the plan and run the skill again. This converts the interrogation into an operating checklist for leadership, legal, data engineering, product, and finance teams.
Extend the skill for your organization
Teams can improve the cdo-review skill by adding company-specific governance rules: approved data classes, consent standards, retention policies, vendor review thresholds, model-training restrictions, and data-product approval gates. Keep those additions separate from the core skill text so the original CDO forcing-question style remains clear while your local constraints make the output more actionable.