| name | rz-website-audit |
| description | Use this skill whenever Riché invokes /rz-website-audit, asks to "run the website audit", "run the weekly site review", "audit richezamor.com", or when the Cowork shortcut fires the Sunday 8pm trigger with the body "Run the rz-website-audit skill." Also invoke for any request that mentions website audit, website QA, website scorecard, AI citation tracking, website regression check, weekly site review of richezamor.com, SEO health check, AIO health check, or competitor benchmarking against richezamor.com. This is the orchestrating skill that runs the full weekly audit across SEO (S1–S7 atomic), AIO (A1–A7 atomic), Traffic & Engagement, Usability, Design, Brand, Technical QA, the Ask Riché chatbot, Keyword Research, and Competitor Benchmarking. Produces a Notion page in the Weekly Website Audits database, creates up to 5 P0/P1 Linear tasks in the Brand project with spaced due dates Mon–Fri, and pings #brand in Slack with a one-line traffic-light headline.
|
Website Audit — Riché Zamor
Weekly site audit of richezamor.com. Runs every Sunday at 8pm America/New_York via Claude Cowork shortcut, or manually via /rz-website-audit in Claude Code. Produces a structured Notion page, issues up to 5 prioritized Linear tasks, and posts a single-line headline to Slack.
Before beginning work
The corpus has migrated to Notion. Notion is the source of truth — the local corpus/* paths cited below are historical and may drift. Always load these references via the Notion MCP (mcp__bc2cd475-c3cd-49fa-a4ab-02ee9f795171__notion-fetch) before executing this skill:
- Strategy Stack README — Notion page
357ac0ea-4f65-81b8-98b4-ffd0f376198c (Brand > README — Strategy Stack). Doc ownership, canonical constants (richezamor.com is the acquisition channel via organic + AIO), and the Decision-of-Record log. Honor it; flag drift.
- Your sections of the Corpus at
Projects > RZ Claude Skills > Corpus:
website-audit → page 357ac0ea-4f65-8158-be44-e9f5ccfd8119 (canonical: methodology, dimensions SEO/AIO/categories, competitor benchmarking, databases)growth → page 357ac0ea-4f65-812a-a480-d3b7ab463bc2 (SEO methodology owned by rz-growth-marketing, channels positioning, target keywords)voice → page 357ac0ea-4f65-8194-ae1e-e5147adad60c (brand hygiene checks must reference voice canon)
Each Corpus directory page lists its child entries. Fetch only the specific entries you need.
Cross-skill dependencies (read these first)
The audit is a tactical skill. It does not own SEO methodology or brand voice. It LEVERAGES the canonical sources owned by other role-based skills.
SEO and keyword research are owned by rz-growth-marketing. Pull from corpus/growth/seo/ for keyword research methodology, SERP review protocol, topic clusters, the monthly Keyword Planner workflow, and the Target Keywords DB schema. The audit's S1–S7 atomic dimensions and K1–K5 keyword-research category dimensions reference this corpus; they do NOT redefine SEO concepts.
GSC data is sourced from BigQuery, not Ahrefs. The audit reads the GSC bulk export at rz-analytics-495304.searchconsole.* (tables: searchdata_url_impression, searchdata_site_impression, ExportLog). The export is forward-only: ~48 hours after Search Console's bulk-export setting was enabled, daily snapshots arrive; there is no historical backfill. S1/S3/K2 require ~28 days of accumulated data before they can fire meaningfully. Anonymized rows in the export are URL-only — query strings are not retained — so anonymized-keyword recovery is out of scope at this stack tier.
Brand voice is owned by rz-copywriting. Pull from corpus/voice/ for the fatal-fifteen AI-tells list, voice anti-patterns, terminology rules, and the 50/30/20 domain balance frame. The audit's B1–B5 brand checks are detection rules ON TOP of the voice canon; they do NOT redefine voice rules.
If a dimension entry's text contradicts these source-of-truth corpora, the source-of-truth corpora win. Surface the contradiction in rz-self-improve for cleanup.
Load from corpus
Read these in order before running. The corpus carries the actual audit logic; this file is the orchestrator.
Methodology (always)
corpus/website-audit/methodology/bootstrap.mdcorpus/website-audit/methodology/parallel-data-collection.mdcorpus/website-audit/methodology/severity-scoring.mdcorpus/website-audit/methodology/report-assembly.mdcorpus/website-audit/methodology/task-issuance.mdcorpus/website-audit/methodology/slack-notification.md
Database schemas (always)
corpus/growth/seo/target-keywords-schema.md (owned by rz-growth-marketing)corpus/website-audit/databases/competitors-schema.mdcorpus/website-audit/databases/weekly-audits-schema.md
SEO atomic dimensions
corpus/website-audit/dimensions/seo/s1-ranking-regression.mdcorpus/website-audit/dimensions/seo/s2-low-ctr.mdcorpus/website-audit/dimensions/seo/s3-quick-wins.mdcorpus/website-audit/dimensions/seo/s4-indexability.mdcorpus/website-audit/dimensions/seo/s5-metadata-completeness.mdcorpus/website-audit/dimensions/seo/s6-structured-data.mdcorpus/website-audit/dimensions/seo/s7-internal-linking.md
AIO atomic dimensions
corpus/website-audit/dimensions/aio/a1-citation-absence.mdcorpus/website-audit/dimensions/aio/a2-citation-rank-decline.mdcorpus/website-audit/dimensions/aio/a3-crawler-accessibility.mdcorpus/website-audit/dimensions/aio/a4-js-only-content.mdcorpus/website-audit/dimensions/aio/a5-quotability-density.mdcorpus/website-audit/dimensions/aio/a6-entity-consistency.mdcorpus/website-audit/dimensions/aio/a7-ai-friendly-schema.md
Category-level dimensions
corpus/website-audit/dimensions/categories/traffic-engagement.mdcorpus/website-audit/dimensions/categories/usability.mdcorpus/website-audit/dimensions/categories/design.mdcorpus/website-audit/dimensions/categories/brand.mdcorpus/website-audit/dimensions/categories/technical-qa.mdcorpus/website-audit/dimensions/categories/chatbot.mdcorpus/website-audit/dimensions/categories/keyword-research.md
Keyword research methodology (owned by rz-growth-marketing)
corpus/growth/seo/free-stack-overview.mdcorpus/growth/seo/keyword-planner-monthly.mdcorpus/growth/seo/serp-review-protocol.mdcorpus/growth/seo/topic-clusters.md
Competitor benchmarking
corpus/website-audit/competitor-benchmarking/read-protocol.mdcorpus/website-audit/competitor-benchmarking/what-gets-benchmarked.md
Constants
WEBSITE_DOMAIN = richezamor.com
MAX_WEEKLY_TASKS = 5
RUN_TIMEZONE = America/New_York
VERCEL_PROJECT_ID = prj_qtI2I2fvlAH5qeTnc4EVW52stG9t
VERCEL_TEAM_ID = team_G5bsJ43GY5r9RqMjT4iEYO5c
LINEAR_TEAM_ID = 72132418-b477-4450-a30a-77391d5cfc47
LINEAR_PROJECT_ID = 085484ef-b523-4142-bce2-7f9a23a05fa1
LINEAR_ASSIGNEE_ID = 646e8aef-65f5-47de-ba40-4f2ad99bbc15
NOTION_PARENT_ID = 333ac0ea4f658086bcafcd3f53299b89
NOTION_AUDIT_DB_ID = 6d5673aef5a94f17867602778949f869
NOTION_AUDIT_DS_ID = 889093b4-fb7d-4db5-a871-19e7bbe159fd
TARGET_KEYWORDS_DB_ID = f5521fe72dec42c1808555fbac2d0d62
TARGET_KEYWORDS_DS_ID = df1d8efa-ff31-4b18-beb5-0998e9dd9bc2
COMPETITORS_DB_ID = c99eaa6d2e7448f0859ce6feba22a3ac
COMPETITORS_DS_ID = 596ee7d6-46d8-464a-adfc-489576de0052
CONTENT_STRATEGY_PAGE_ID = 333ac0ea4f6581518651d730d706e017
CONTENT_TOPICS_DB_ID = 2c224f29490d44f99e7b58cb2e377086
SLACK_CHANNEL = #brand
DOMAIN_BALANCE_TARGET = 50% Context Layer / 30% PM / 20% Leadership
KEYWORD_PLANNER_STALE_DAYS = 30
SERP_ANALYSIS_TOP_N = 5
QUARTERLY_REBASELINE_MONTHS = January, April, July, October (first Sunday)
Process
Step 1 — Bootstrap
Per methodology/bootstrap.md:
- Verify required MCP connectors are reachable (Notion, Linear, Slack, Vercel, Tavily, BigQuery).
- Read prior audit page from Weekly Audits DS for diff baseline.
- Read Target Keywords DB and Competitors DB into memory.
- Initialize empty findings buffer for each dimension.
Step 2 — Parallel data collection
Per methodology/parallel-data-collection.md. Concurrent fetches:
- GSC data via BigQuery bulk export at
rz-analytics-495304.searchconsole.* (positions, CTR, impressions, anonymized URL-level rows)
- richezamor.com crawl: every page in sitemap, fetch + render-without-JS check + indexability heuristics (HTTP status, robots.txt allowance, canonical correctness, orphan detection — feeds S4)
- AIO 20-query set across the LLMs Riché tracks
- Competitors DB benchmarks per
competitor-benchmarking/read-protocol.md (~30 min, the longest block)
- Vercel deployment status and recent build logs
Step 3 — Run dimensions
For each dimension corpus entry, evaluate against collected data. Write findings to the buffer with severity (P0/P1/P2) per methodology/severity-scoring.md. Order:
- SEO S1–S7 (atomic — each is its own corpus entry)
- AIO A1–A7 (atomic)
- Traffic & Engagement T1–T5 (single corpus entry)
- Usability U1–U4 (single corpus entry)
- Design D1–D5 (single corpus entry)
- Brand B1–B5 (single corpus entry)
- Technical QA Q1–Q9 (single corpus entry)
- Chatbot C1–C5 (single corpus entry)
- Keyword Research K1–K5 (single corpus entry; SERP review per
serp-review-protocol.md runs here)
- Competitor Benchmarking deltas (writes back to Competitors DB then surfaces deltas)
Step 4 — Assemble report
Per methodology/report-assembly.md. Build the 15-section audit page in memory. Compute headline color per the rules in databases/weekly-audits-schema.md. Do not write to Notion yet.
Step 5 — Issue tasks
Per methodology/task-issuance.md. Take the top P0 + P1 findings, capped at MAX_WEEKLY_TASKS (5). Create Linear issues in LINEAR_PROJECT_ID, assigned to LINEAR_ASSIGNEE_ID, with due dates spaced Mon–Fri across the coming week. Use the create-then-update pattern (create simple, then update with full content) — direct create-with-full-payload errors out unreliably. Capture the TRZ-### IDs.
Step 6 — Write Notion page
Single write to Weekly Audits DS using NOTION_AUDIT_DS_ID. Include the Linear TRZ-### IDs in the Tasks Issued section. Set headline color last, after task issuance is confirmed.
Step 7 — Slack notification
Per methodology/slack-notification.md. One-line headline to SLACK_CHANNEL with traffic-light emoji, task count, and a link to the Notion audit page.
Step 8 — Update Target Keywords DB
Write back Current Position and Last Checked for every Researching/Targeting/Ranking row processed. Fire status promotions (Ranking → Won) per the 4-week confirmation rule in corpus/growth/seo/target-keywords-schema.md.
Step 9 — Update Competitors DB
Already handled in Step 2's competitor block, but confirm Last Audited updated on every row touched.
Quarterly rebaseline
On the first Sunday of January, April, July, and October, the audit additionally:
- Re-scores all 14 competitors regardless of tier
- Reviews Won and Deprioritized rows in Target Keywords DB for any needed status changes
- Logs a quarterly summary section at the top of the audit page
The constant QUARTERLY_REBASELINE_MONTHS flags this; bootstrap reads it.
What this skill does NOT do
- It does not draft or publish content. K3 fires issue Linear tasks; Riché writes the article via
/rz-draft-content.
- It does not refresh Keyword Planner volumes. That's Riché's monthly manual workflow per
corpus/growth/seo/keyword-planner-monthly.md.
- It does not re-tier competitors. Tier changes are a quarterly Riché decision.
- It does not auto-fix any issues it finds. It diagnoses, scores, and issues tasks. Fixes happen via separate skills or manual Riché work.
- It does not run Lighthouse via the official Lighthouse CLI. It approximates Lighthouse scoring from public-web signals only, weighted 40/20/20/20 per
competitor-benchmarking/what-gets-benchmarked.md.
Cross-skill connections
Upstream (the audit reads from these for canonical knowledge):
rz-growth-marketing — owns SEO and keyword research methodology. The audit reads corpus/growth/seo/ for the free-stack approach, SERP review protocol, topic clusters, monthly Keyword Planner workflow, and the Target Keywords DB schema. The audit's K1–K5 dimensions and S1–S7 atomic fires use these as source of truth.rz-copywriting — owns brand voice. The audit reads corpus/voice/ for the fatal-fifteen AI tells and the 50/30/20 domain balance. B1–B5 dimensions detect drift against this canon.
Downstream (the audit hands off to these for fixes):
rz-content-optimize — when K3 fires and a Linear task is issued, that task references rz-content-optimize for keyword + entity coverage when Riché drafts the article.rz-copywriting — B1 (voice) findings reference rz-copywriting's voice rules. The audit detects voice drift; rz-copywriting fixes it.rz-graphic-design — D5 (OG images) findings reference rz-graphic-design when an article ships without proper OG asset.rz-self-improve — runs as the SessionEnd hook in retrospective mode; logs any audit-time exceptions for next-run improvement.
