Google Cloud Cost Optimization Advisor
Generates actionable Google Cloud cost guidance using the Well-Architected Framework.
The skill only generates guidance and does not declare cloud mutations, required permissions, or data exfiltration; no red-line risk is visible. However, it lacks handling guidance for sensitive billing/workload data, user confirmation, rollback, and data-flow disclosure, so points are deducted.
The principles, product examples, assessment questions, and checklist are internally coherent and make the happy path understandable. There is no input schema, abnormal-input handling, dependency guidance, failure feedback, or test evidence, so the static score is conservative.
The target scenario and audience are reasonably clear for Google Cloud cost-optimization assessment. Non-fit boundaries, trigger precision, expected output format, and Chinese-language support are unspecified; referenced Google documentation may also be difficult to reach from mainland China, warranting a deduction.
The documentation is readable and layered, with metadata, repository installation guidance, an Apache-2.0 license, and support channels. The skill lacks output examples, FAQs, known limitations, versioning, changelog information, and explicit maintenance ownership, so points are deducted.
The question set and checklist can support an initial cost review and provide practical value. However, the guidance is generic, lacks a workload-specific output structure and validation procedure, and uses potentially over-broad requirements such as 100% labeling and monthly actions; the static score remains below the execution-dependent ceiling.
The skill includes links to Google Cloud WAF grounding documents, providing limited traceability for its principles. It contains no committed tests, real execution evidence, cross-source corroboration, or independent reproduction, so only limited credit is justified.
- This is a static document review; the skill and its referenced documents were not executed or independently tested.
- Billing, utilization, and organizational cost data may be sensitive; confirm scope, least privilege, and data handling before use.
- Requirements such as 100% labeling coverage and monthly reviews should be adapted to organizational constraints rather than treated as universal hard rules.
- If the runtime cannot reach the referenced Google Cloud documentation, the grounding material may not be readily verifiable.
What does this skill do, and when should you use it?
This skill provides Google Cloud workload guidance based on the Cost Optimization pillar of the Well-Architected Framework. It covers the Inform, Optimize, and Operate phases of FinOps, along with four principles: aligning spending with business value, fostering cost awareness, optimizing resource usage, and optimizing continuously. It gives standard recommendations before asking workload-assessment questions. A validation checklist covers attribution, visibility, budgets, rightsizing, commitments, idle resources, storage, and network egress.
For a Google Cloud service or workload scenario, it generates a comprehensive set of standard cost recommendations covering Cloud Billing reports, BigQuery billing export, Looker Studio, budgets and alerts, Storage Insights, Recommender, Active Assist, FinOps hub, Cloud Hub Optimization, Billing quotas, CUDs, SUDs, Spot VMs, storage lifecycle policies, network tiers, CDN, Interconnect, Direct Peering, managed services, database configuration, and organizational governance. It can then ask workload-assessment questions and apply the supplied checklist to review labels, billing visibility, budgets, rightsizing, commitment strategy, idle resources, storage tiers, and network egress.
- A Google Cloud architect reviewing a production workload against the Well-Architected Framework.
- A FinOps team managing costs across multiple projects or departments and designing attribution, budgets, dashboards, and recurring reviews.
- A platform team running Compute Engine workloads and evaluating rightsizing, CUDs, SUDs, and Spot VMs.
- An engineering team managing Cloud Storage or multi-region architecture and reviewing access patterns, lifecycle policies, and egress costs.
- A technical lead choosing database, serverless, or managed-service configurations while balancing operating cost, capacity, and high availability.
What are this skill's strengths and limitations?
- Covers visibility, resource usage, discounts, storage, networking, databases, and governance.
- Requires complete standard recommendations before follow-up assessment questions.
- Includes concrete Google Cloud products, discount types, and recurring validation checks.
- Distinguishes native Cloud Billing reporting from customized Looker Studio reporting.
- It is guidance and a checklist, not an automation script that changes Google Cloud resources.
- The source provides no test suite, sample outputs, or platform compatibility tests.
- Useful recommendations depend on workload, billing, and organizational context supplied by the user.
- It covers this cost-optimization skill only, not the capabilities of the repository’s other 89 skills.
How do you install this skill?
The skill is located at skills/cloud/google-cloud-waf-cost-optimization/ in the repository. To install the collection, run: npx skills add google/skills. The installer allows selection of specific skills. The README does not provide a separate command dedicated only to this skill.
How do you use this skill?
Enable the skill in an Agent Skills-compatible client and describe the Google Cloud workload and cost objective. Example trigger: “Assess this production workload using the Cost Optimization pillar of the Google Cloud Well-Architected Framework. It uses Compute Engine, Cloud Storage, and Cloud SQL; provide complete recommendations and a validation checklist.” The source does not document client-specific enablement or triggering syntax.
How does this skill compare with similar options?
The source distinguishes Cloud Billing reports from Looker Studio: Cloud Billing reports provide standard spending views in the console, while Looker Studio supports advanced, shareable, customized reporting. It also distinguishes CUDs from SUDs: CUDs suit predictable steady-state workloads, while SUDs are passive discounts for instances running for a significant portion of the month without commitments.