SEO Audit workflow skill. Use this skill when the user needs to diagnose and audit SEO issues affecting crawlability, indexation, rankings, and organic performance while preserving upstream workflow intent, copied support files, and provenance.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
SEO Audit workflow skill. Use this skill when the user needs to diagnose and audit SEO issues affecting crawlability, indexation, rankings, and organic performance while preserving upstream workflow intent, copied support files, and provenance.
This skill preserves the intent of the upstream SEO audit workflow while making it more reliable to execute as a disciplined diagnostic review.
Use it when the goal is to identify, explain, and prioritize SEO issues affecting crawlability, indexation, rendering, canonicalization, structured data, content quality, and organic performance. Outputs should be evidence-led, scoped, and decision-oriented.
This is an audit skill, not an implementation skill.
Diagnose before recommending fixes.
Separate confirmed technical issues from ranking hypotheses.
Do not claim penalties, algorithm hits, or causation without direct evidence.
Prioritize by business impact, scope, and confidence.
Preserve upstream workflow intent and provenance when handing off or merging findings.
When to Use
Use this skill when the user asks for any of the following:
A technical SEO audit of a site, section, template, or migration.
Investigation of crawlability, indexation, canonicalization, rendering, or sitemap problems.
Review of structured data eligibility, page experience signals, or mobile usability concerns.
Diagnosis of organic traffic loss where technical causes are plausible.
A severity-ranked findings report with evidence, likely cause, and next validation step.
Do not use this skill as the primary workflow when:
The user only wants copywriting, metadata writing, or content generation.
The task is pure implementation of fixes with no audit or diagnosis phase.
The user wants guarantees about rankings, penalties, or recovery timelines.
Access is too limited to support meaningful conclusions and the user has not accepted a constrained audit.
Scope Gate
Before starting, confirm the operating boundary. Ask for missing inputs first.
Target scope: domain, subdomain, folder, country/language variant, or page set.
Business context: key templates, critical pages, recent migrations, traffic drop timing, release history.
If core access is missing, continue only with an explicit limitation statement such as:
"This is a constrained SEO audit based on public evidence and supplied artifacts only. Findings may miss issues that require Search Console, logs, or authenticated crawling data."
Required Output Format
Return findings in a structure that supports action and review:
Audit scope and limitations
Executive summary
Top issues by severity
Detailed findings with:
issue title
severity
affected scope
evidence
likely cause
SEO impact
confidence level
recommended next validation step
Open questions / missing access
Priority order: fix now, validate next, monitor
A useful finding is specific enough that another operator can verify it without guessing.
Workflow
Confirm scope and access
Define what is in scope and what is out.
Record available evidence sources.
Note whether the audit is full, sample-based, or hypothesis-led.
Collect baseline evidence
Capture representative URLs by template and business importance.
Gather robots.txt, sitemap locations, canonical targets, response behavior, rendered HTML evidence, and Search Console signals if available.
Note historical context such as redesigns, migrations, robots changes, or JavaScript framework changes.
Audit crawl and index controls
Check whether important URLs are discoverable through links, sitemaps, and internal architecture.
Performance and mobile usability: issues that materially affect crawl, rendering, or user access.
Severity Model
Use a simple and defensible severity scale:
Critical: prevents crawl, indexing, or correct canonical selection for high-value pages or large site sections.
High: materially suppresses discoverability, indexing quality, or search appearance across important templates.
Medium: meaningful quality issue with moderate scope or uncertain impact.
Low: localized issue, hygiene problem, or optimization opportunity.
Monitor: no confirmed defect yet; track with a validation method.
Severity should reflect business impact plus scope, not just technical purity.
Examples
Example 1: Good finding
Issue: Product pages excluded because canonical points to category URLs
Severity: High
Affected scope: ~3,200 product detail pages
Evidence:
- Sample product URLs return 200 and are internally linked
- HTML canonical points to parent category URL
- Search Console coverage for sampled pages shows alternate page with proper canonical tag
Likely cause: Template-level canonical logic inherited category canonical setting
SEO impact: Product pages are unlikely to be indexed as independent landing pages
Confidence: High
Recommended next validation step: Confirm the canonical tag rule across templates and sample recrawl after correction
Example 2: Weak finding to avoid
Issue: Google may be penalizing the site
Severity: Critical
Evidence: Traffic went down
Why weak: No direct evidence, no technical mechanism, no separation between correlation and cause
Do start with scope, access, and known business events before reviewing symptoms.
Do use representative URL samples across templates instead of relying on one page.
Do distinguish public-web evidence from Search Console-only evidence.
Do state limitations when access gaps reduce confidence.
Do tie every major finding to proof, affected scope, and next validation step.
Do identify whether an issue is page-level, template-level, or architecture-level.
Do prioritize blockers before metadata refinements.
Do not promise ranking lifts, recovery timelines, or penalty explanations without direct evidence.
Do not mark every issue High; calibrate by impact and breadth.
Do not confuse canonical preference with indexation guarantee.
Do not treat all Core Web Vitals or performance issues as primary SEO blockers if crawl/index signals are healthy.
Do not recommend broad robots blocking or noindex changes without clearly stating the risk.
Do not claim JavaScript SEO failure unless raw/render evidence shows missing or altered critical signals.
Troubleshooting
Symptoms: Many important URLs are absent from search results, but pages load normally in the browser.
Solution: Check indexation evidence first: noindex directives, canonicals, redirects, duplicate clusters, sitemap inclusion, and Search Console page indexing states. A page being reachable is not the same as being indexable.
Symptoms: Search Console shows "Crawled - currently not indexed" on many pages.
Solution: Avoid a single-cause explanation. Review duplication, thin template content, weak internal linking, soft-404 characteristics, and canonical ambiguity. Treat this as a diagnostic state, not a verdict.
Symptoms: Important content exists after user interaction but not reliably in initial HTML.
Solution: Compare raw versus rendered output for content, links, canonicals, metadata, and structured data. If critical elements depend on delayed JS execution, classify it as a rendering risk and scope the affected templates.
Symptoms: Sitemaps exist, but discovery and indexing remain poor.
Solution: Check whether sitemap URLs are canonical, indexable, return valid status codes, and align with internal linking. Sitemaps support discovery; they do not override poor canonical or quality signals.
Symptoms: Structured data is present but rich result eligibility is inconsistent.
Solution: Validate syntax, required properties, policy fit, and consistency with visible content. Separate markup validity from search feature eligibility.
Symptoms: Organic traffic dropped after a migration or redesign.
Solution: Compare pre/post behavior for redirects, canonicals, status codes, internal links, robots rules, sitemap coverage, and rendering changes. Prioritize migration regressions before broader content theories.
Additional Resources
references/review-criteria.md — open this during the audit to classify issue families, evidence standards, severity logic, and validation paths consistently.
examples/review-example.md — open this when you need a concrete example of a scoped, evidence-based SEO audit write-up.
Related Skills
No related local skills were supplied in the source context.
