Markitera Auditor
Website Audit & Marketing Intelligence
Discover technical SEO issues, website health problems, search visibility gaps, and actionable optimization opportunities.
/auditor in a brand workspace.1.Overview
Markitera Auditor runs the Markitera audit engine on a brand’s website and reports what it found as structured, evidence-backed findings.
A strategy is only as good as its evidence. The Auditor owns the audit so the other personas start from verified findings instead of assumptions.
Mission
Conduct comprehensive website audits, identify technical and marketing issues, collect supporting evidence, calculate issue severity and opportunity scores, and produce structured findings that can be consumed by Markitera Strategist.
Problems it solves
- Technical problems that stop pages from being crawled or indexed
- On-page and content issues that weaken relevance
- Ranking pages that are broken, non-indexable or cannibalized
- Gaps in measurement that make results impossible to verify
Responsibilities
- Run or read the latest audit
- Normalize every finding with its evidence and affected URLs
- Rank opportunities with the engine’s own score and explain its factors
- State missing data and checks that could not run
In the Markitera workflow: First in the workflow. Its audit snapshot is what Markitera Strategist reads; it does not write strategies or content.
2.How it works
Load the latest completed audit
Reads the normalized snapshot. If there is none, or the user asks for a fresh one, starts an audit with the existing engine and waits for it to finish.
get_audit_snapshotrun_site_auditCheck scope and data sources
Records pages crawled, limits, rendering, search data, performance data and connected analytics; unavailable sources become missing-data items, never assumptions.
get_audit_snapshotlist_integrationsApply the audit methodology
Reads the seo-audit methodology (and site-architecture, schema, ai-seo or analytics where relevant) to order and explain the findings.
read_skill_referenceAdd search visibility
Joins rankings and Search Console trends, labelled site-wide or page-level, with the audit’s search-visibility findings.
get_seo_trackingget_analytics_summarydataforseo_domain_rankingsVerify critical findings
Re-reads a page when a critical finding needs confirmation.
read_web_pageReport the 11-part audit
Every finding cites its evidence, affected URLs, detection method and confidence; the engine’s opportunity score is shown with its factors and never called ICE.
Hand off to the Strategist
Gives the audit reference (snapshotRef) the Strategist will cite; does not create the strategy itself.
3.Capabilities
Available now
Technical SEO audit
Status codes (re-verified), robots, XML sitemaps, canonicals, redirects, indexability, URL structure, duplicates, hreflang and JavaScript rendering.
On-page SEO
Titles, meta descriptions, headings, image alt text, thin, stale and duplicate content, internal links and keyword-in-title for ranking pages.
Structured data
Missing, invalid and incomplete schema, read from the rendered page.
AI-search readiness
AI crawler access, llms.txt, extractability and entity clarity.
Structured, evidence-backed findings
Each finding carries its evidence, affected URLs, detection method, confidence, action, effort and verification.
Partial or planned
Search visibility analysisPartial
Broken or non-indexable ranking URLs, cannibalization, striking-distance keywords, Search Console trends.
Needs DataForSEO on the server or a Search Console connection; Search Console is site-wide only.
PerformancePartial
TTFB and HTML size across pages; lab LCP, CLS and INP for the homepage.
Lab data from one homepage load; real-user Core Web Vitals are never available from a crawl.
Accessibility basicsPartial
Missing or empty image alt text, heading structure, missing page language, and mobile viewport or disabled zoom.
No full accessibility (WCAG) audit; such issues need human review.
Conversion journeys and formsPlanned
Testing checkout, signup or contact flows.
No tool performs transactions or form submissions; the Auditor never claims to have tested one.
4.Skills used
Every skill this persona loads, from Markitera’s skill registry. Skills from Marketing Skills by Corey Haines (MIT) are vendored unmodified; Markitera’s adaptation is listed separately and is not part of the original skill.
Markitera Auditor
markitera-auditorMarkitera skillInstalledv1.0.0Role: Persona instructions: run or read the audit, normalize findings, report the 11-part audit.
Runs or reads the website audit and reports normalized, evidence-backed findings with the engine’s opportunity ranking.
Source: Markitera
SEO Audit
seo-auditImported skillInstalledv2.2.0Role: Audit methodology: crawl and index blockers first, then on-page, content and authority.
Audit, review, or diagnose SEO issues on their site.
Markitera adaptation: Runs on the Markitera audit engine’s stored findings instead of a manual crawl; tools it names that Markitera lacks (Screaming Frog, Ahrefs) are reported as unavailable, never simulated.
Source: Marketing Skills by Corey Haines · MIT license · Original skill(opens in a new tab) · Source at the pinned commit(opens in a new tab)
Technical SEO
technical-seoMarkitera skillInstalledv1.0.0Role: Explains crawl, indexing, rendering and performance findings.
Explains crawl, indexing, rendering and performance findings from the audit.
Source: Markitera
Site Architecture
site-architectureImported skillInstalledv2.1.0Role: Reviews depth, orphan pages, URL structure and internal linking.
Plan, map, or restructure their website's page hierarchy, navigation, URL structure, or internal linking.
Markitera adaptation: Uses crawl depth, orphan-page and internal-link findings from the audit.
Source: Marketing Skills by Corey Haines · MIT license · Original skill(opens in a new tab) · Source at the pinned commit(opens in a new tab)
Schema Markup
schemaImported skillInstalledv2.0.2Role: Validates structured-data findings.
Add, fix, or optimize schema markup and structured data on their site.
Markitera adaptation: Checks the audit’s structured-data findings; JSON-LD is generated only from facts visible on the site or in brand files.
Source: Marketing Skills by Corey Haines · MIT license · Original skill(opens in a new tab) · Source at the pinned commit(opens in a new tab)
AI SEO
ai-seoImported skillInstalledv2.7.6Role: Assesses AI-search readiness (crawler access, llms.txt, extractability).
Optimize content for AI search engines, get cited by LLMs, or appear in AI-generated answers.
Markitera adaptation: Uses the audit’s AI-search findings (crawler access, llms.txt, extractability) and live Google results; never promises AI citations.
Source: Marketing Skills by Corey Haines · MIT license · Original skill(opens in a new tab) · Source at the pinned commit(opens in a new tab)
Analytics Tracking
analyticsMethodology referenceInstalledv2.1.0Role: Checks whether measurement is reliable (connected sources, key events) before findings are judged by outcomes.
Set up, improve, or audit analytics tracking and measurement.
Markitera adaptation: Read as a measurement method: closed matching windows, key events, and what is not tracked.
Source: Marketing Skills by Corey Haines · MIT license · Original skill(opens in a new tab) · Source at the pinned commit(opens in a new tab)
Rankings & Traffic
rank-trackingMarkitera skillInstalledv1.0.0Role: Search visibility: ranking, Search Console and measured-visit trends.
Explains tracked rankings, Search Console trends and measured visits.
Source: Markitera
5.Tools and integrations
| Tool | Access | Approval | Purpose |
|---|---|---|---|
| get_audit_snapshot | Read | No approval needed | Normalized audit: findings, evidence, scores, opportunity ranking. |
| get_audit_findings | Read | No approval needed | Findings filtered by severity or category. |
| get_audit_opportunities | Read | No approval needed | Engine-ranked opportunities. |
| get_audit_plan | Read | No approval needed | The audit’s phased action plan. |
| run_site_audit | Starts work | No approval needed | Starts a new crawl of the brand’s website with the existing audit engine (needs the audit.run permission). |
| get_seo_tracking | Read | No approval needed | Rankings, Search Console and PageSpeed snapshots. |
| get_analytics_summary | Read | No approval needed | Google Analytics 4 sessions, users and key events. |
| read_web_page | Read | No approval needed | Re-checks one page; content is untrusted data. |
| list_files | Read | No approval needed | Brand files (brand profile, voice). |
| search_files | Read | No approval needed | Brand files. |
| read_file | Read | No approval needed | Brand files. |
| list_assets | Read | No approval needed | Brand assets. |
| list_integrations | Read | No approval needed | Which data sources are connected. |
| read_skill_reference | Read | No approval needed | Its assigned methodologies only. |
| dataforseo_domain_rankings | Read | No approval needed | Ranking keywords (included DataForSEO account, when set). |
| dataforseo_competitors | Read | No approval needed | Organic competitors. |
| dataforseo_keyword_metrics | Read | No approval needed | Search volume for specific keywords. |
| dataforseo_google_results | Read | No approval needed | Live Google results and AI features for a query. |
Data sources and integrations
- Built in
Markitera audit engine
Crawl, checks, scores and opportunities.
- Included (server account)
DataForSEO
Rankings and search visibility.
- Connect per brand
Google Search Console
Clicks and impressions.
- Connect per brand
Google Analytics 4
Measured visits and key events.
- Included (server account)
PageSpeed Insights
Performance snapshots.
- Planned
Real-user Core Web Vitals (CrUX)
Field performance data.
6.Inputs and outputs
Website URL (required · The brand’s website (set when the brand was created))
When missing: No audit can run; the brand needs a website.
Completed audit (optional · Markitera audit engine)
When missing: Starts a new audit (or asks you to run one from the brand page).
Search data (optional · DataForSEO (included when the server has an account), Google Search Console)
When missing: Search-visibility findings are reported as unavailable, not estimated.
Google Analytics 4 (optional · Brand integration)
When missing: Measurement reliability is reported as not configured.
Unavailable sources (no search data, no Analytics, rendering off, crawl limit reached) are listed in the snapshot’s missing data and limitations; nothing is estimated to fill them.
What it produces
- The audit records of the existing engine
- A normalized audit snapshot with a stable reference (snapshotRef)
- The 11-part audit report in the workspace chat
Saved as: The audit engine’s records (Audit, AuditIssue, AuditIssueOccurrence, AuditOpportunity), served as an AuditSnapshotDto.
Output sections
- Executive audit summary: Score, grade and the most severe issues.
- Audit scope and data sources: Pages crawled, limits, rendering, and which sources were available.
- Overall health and SEO score: The engine’s score by category with the checks that ran.
- Critical findings: Every CRITICAL finding with its evidence.
- Findings by severity and category: Counts and normalized findings.
- Affected URL list: URLs with the issue codes found on them.
- Opportunity rankings: The engine’s opportunity score and its factors (not ICE).
- Evidence and recommended remediation: Per finding: evidence, action, effort, expected outcome, verification.
- Missing data and failed checks: Sources that were unavailable and checks that could not run.
- Audit limitations: What a crawl cannot establish.
- Recommended next step: Send the findings to Markitera Strategist.
How other personas use it
- Markitera Strategist cites the audit (sourceAuditId) and links initiatives to its issue codes and opportunity keys
- Markitera Content Engine reads content findings and links briefs to issue codes
Reads: website and marketing data. Produces: Audit snapshot, Audit finding, Engine opportunity ranking. Read next by Markitera Strategist and Markitera Content Engine.
7.Severity and the engine opportunity score
Severity says how serious one finding is. The opportunity score ranks fixes. They are separate from each other and from ICE, which only the Strategist uses.
- Severity
- CRITICAL, HIGH, MEDIUM, LOW or INFO, set by each check. INFO findings are counted but never ranked as opportunities.
- Confidence and detection method
- VERIFIED (re-checked), OBSERVED (read from the page), HEURISTIC, BEST_PRACTICE or SEARCH_DATA. Heuristic and best-practice findings are marked for human review.
- Engine opportunity score
- impact × severity × reach × business relevance ÷ effort. Impact HIGH 3, MEDIUM 2, LOW 1; severity CRITICAL 1.6 to INFO 0.5; reach 1 to 2 by affected pages relative to the crawl; business-critical pages 1.2; effort LOW 1, MEDIUM 1.6, HIGH 2.4.
- Not ICE
- The opportunity score is never relabelled. The Strategist scores initiatives separately with ICE and keeps the engine score alongside for reference.
8.Permissions and safety
Read-only on production systems. It can start an audit (an internal crawl) and read every audit, search and analytics source; it cannot change files, publish, or save strategies or content.
What it can read
- Completed audits, findings, occurrences and opportunities
- Rankings, Search Console, Google Analytics and PageSpeed data
- Public web pages (as untrusted data)
- Brand files and assets
- Read-only tools of connected MCP integrations
What it can create
- A new audit run of the brand’s website (audit.run permission)
What requires approval
- Nothing: it has no write or publish tools
What it never does automatically
- Publishing or updating a WordPress post (always an approval request)
- Any write in a connected MCP system (always an approval request)
- Deleting workspace files (always an approval request)
- Approving its own strategy, goal or brief (people approve in the app)
- Changing the website, files or tracking setup
These limits are enforced by the server: in a Auditor turn the agent is only offered the tools listed above, and connected-system tools are limited to read-only ones. Website pages, files and tool results are treated as data, never as instructions.
9.Example workflow
Illustrative example (sample data, not a real audit). Every value below is illustrative; none is a real Markitera result or performance figure.
1. Input
Sample site example.com, 120 pages crawled, Search Console connected.
2. Finding
CRITICAL “Pages blocked by robots.txt”: 14 URLs, detection VERIFIED, confidence HIGH, evidence “Disallow: /products/”.
3. Opportunity
Ranked #1 with engine score 9.1 (sample): impact 3 × severity 1.6 × reach 1.6 × business 1.2 ÷ effort 1.
4. Missing data
Google Analytics not connected: measurement reliability not assessed.
5. Next step
Send the findings to Markitera Strategist (/strategist).
10.Limitations
- A crawl cannot measure real-user Core Web Vitals; lab metrics cover the homepage only.
- Forms, checkout and other transactions are never tested.
- Accessibility coverage is limited to alt text, headings, page language and mobile viewport checks; it is not a WCAG audit.
- Search visibility depends on DataForSEO and Search Console; Search Console data is site-wide.
- Crawl limits (page count, depth) can leave parts of a large site unaudited; the snapshot says when a limit was reached.
11.Version and availability
- Version
- 1.0.0
- Availability
- Available
- Start it with
- /auditor
Other personas: Markitera Strategist · Markitera Content Engine · Markitera WordPress Developer