check-relevance

$npx mdskill add MicrosoftDocs/cloud-adoption-framework/check-relevance

Assess if technical guidance reflects modern Azure best practices and architectural patterns.

  • Determines if existing documentation recommends outdated or superseded cloud patterns.
  • Utilizes external documentation search to compare old advice against current standards.
  • Compares document intent against live search results to spot strategic drift.
  • Suggests specific, evidence-backed rewrites for major architectural shifts.

SKILL.md

.github/skills/check-relevanceView on GitHub ↗
---
name: check-relevance
description: Assesses whether guidance is still relevant and framed for modern Azure approaches. Use when asked to check if content is still current or strategically relevant.
---

You are a strategic content relevance assessor for Azure technical documentation. Evaluate whether a document's overall guidance still recommends the right approach for today's Azure landscape. Other skills handle service names, typos, links, and security — focus on the big picture.

## Process

1. Identify the document's purpose — what problem/scenario/architecture does it address?
2. Use `microsoft_docs_search` to search for the same topic in current docs (2-3 searches).
3. Compare: Does current guidance still recommend this approach?
4. If changed, suggest specific rewrites backed by MCP evidence.

## What to flag

- Superseded architectures/patterns (e.g., custom NVA hub → Azure Virtual WAN)
- Manual approaches now handled by managed services natively
- Workarounds for limitations that have been resolved
- Outdated paradigm framing (e.g., on-premises-first vs. cloud-native default)

## What NOT to flag

- Service name changes, retired products, links, typos (other skills own these)
- Guidance that's still the current recommended approach
- Content that's sound but could benefit from minor updates

## Rules

- Only flag STRATEGIC issues backed by MCP evidence
- No "(preview)" labels, no absolute URLs, no `/en-us/` locale prefixes
- Rewrite retired approaches naturally — no retirement notices
- Keep replacements focused on the specific outdated framing
- When adding new content, add it at ONE location only — never insert the same paragraph at multiple points
- Before making changes, verify they won't contradict other parts of the document

More from MicrosoftDocs/cloud-adoption-framework

SkillDescription
acronym-checkEnsures abbreviations are expanded on first use per Microsoft style. Use when asked to check acronyms or abbreviations in documentation.
active-voiceRewrites passive voice to active/imperative and enforces second person. Use when asked to fix passive voice or improve writing directness.
api-consistencyEnsures Azure REST API and CLI consistency across a document. Use when asked to check API calls, CLI parameters, or REST consistency.
azure-validationValidates portal navigation paths, detects contradictions against Microsoft Learn documentation, and checks alignment with current Azure implementation standards. Use when asked to verify Azure portal instructions or validate content accuracy.
check-accessibilityChecks image accessibility compliance per Microsoft Learn standards. Use when asked to review images for alt text, lightbox paths, or complex image markup.
cloud-adoptionHow to adopt and integrate the Microsoft Azure cloud into your organization. Strategy, policies, organizational readiness, architecture, platform landing zone, governance, security, health, and operations.
code-expertIdentifies definite errors in code samples including syntax errors, missing imports, and deprecated APIs. Use when asked to review code samples for correctness.
code-freshnessReviews code samples for modern patterns, deprecated APIs, outdated dependencies, and alignment with current Microsoft best practices. Use when asked to update or modernize code samples.
code-verifierVerifies code samples against official Microsoft/Azure examples and syntax rules. Use when asked to verify code correctness or validate API usage.
error-checkIdentifies severe factual errors in content. Use when asked to fact-check documentation for major inaccuracies.