seo-geo-audit

name	seo-geo-audit
description	Run public-safe SEO and GEO audits for any project using local evidence first, then optional current external evidence when available.

SEO/GEO Audit

Use this skill when the user asks for an SEO audit, GEO audit, AI-search visibility review, content opportunity map, answer-engine readiness review, Search Console analysis, crawl analysis, log analysis, or reusable search briefing for any project.

First Principles

• Start from the current project, not a remembered brand or prior audit.
• Treat every project as prelaunch or evidence-limited until local evidence proves otherwise.
• Separate first-party evidence, live external evidence, and inference.
• Prefer entity clarity, answerability, proof, crawlability, internal linking, technical accessibility, and content usefulness over keyword stuffing.
• Keep search overlap separate from direct product competition.
• Never silently treat missing connectors, missing exports, or blocked crawls as poor performance.

Public-Safe Rules

• Do not copy raw credentials, tokens, cookies, private keys, analytics IDs, private URLs, customer records, email addresses, personal data, or internal planning excerpts into the final report.
• Refer to sensitive sources by role and path shape, such as analytics export, access log sample, crawl export, or local planning document.
• Redact sensitive substrings before quoting a file path, URL, header, query string, log line, or environment variable.
• If a finding depends on sensitive evidence, summarize the behavior and cite the evidence category instead of pasting the raw evidence.
• Do not write audit artifacts outside the user-selected project or output directory.
• Do not create project-specific defaults in this skill. Project packs belong in project repositories, not in the public plugin.

Workflow

1. Discover context:
   • Identify the project root, site URL if provided, app framework, generated site output, sitemap, robots file, routes, content directories, docs, and available evidence exports.
   • If the user provides a GitHub URL, inspect the local read-only mount first when available, and use a writable checkout only for edits.
2. Classify evidence:
   • Mark each source as local, connector, live web, or inference.
   • Mark the audit as draft, prelaunch, local-only, or evidence-backed.
3. Check safety:
   • Scan planned citations and report snippets for credentials, customer data, private paths, query strings, and proprietary excerpts.
   • Redact before writing or presenting results.
4. Audit:
   • Technical SEO: crawlability, indexability, canonical URLs, metadata, structured data, sitemap, robots rules, page status, performance signals, and accessibility blockers.
   • Content SEO: topic coverage, page intent, information gain, internal links, proof points, comparisons, glossary needs, and conversion paths.
   • GEO readiness: entity definitions, answerable questions, concise claims, citation-worthy proof, source clarity, author or organization clarity, schema support, and retrieval-friendly page structure.
5. Produce outputs:
   • Keep reports concise and evidence-labeled.
   • Include assumptions and manual next steps when evidence is missing.
   • Save artifacts only when the user asks for files or the workflow requires checked-in deliverables.

Output Contract

Use these sections unless the user requests a different format:

• Executive Summary
• Evidence Used
• Evidence Bundle Readiness
• Scores
• Top Findings
• Priority Backlog
• Page Briefs
• GEO Benchmark
• Assumptions
• Next Steps

Scoring Contract

Use a 0-5 scale when scoring:

• 0: no usable evidence or severe blocker
• 1: present but broken or misleading
• 2: partial, thin, or inconsistent
• 3: adequate baseline
• 4: strong with minor gaps
• 5: best-in-class for the project stage

Always explain when a score is capped by missing evidence.

Next Steps

After each pass, tell the user:

• what was actually inspected
• what was inferred
• which evidence would improve confidence
• what they need to do manually, if anything