The Complete Claude
SEO Playbook

---
name: seed-expander
description: Turn one seed keyword into a full keyword universe with volume and difficulty. Use at the start of any new topic.
---

Take the seed. Expand via modifiers, questions, comparisons, alternatives, and long-tail variants.
For each keyword: volume | difficulty | intent | whether the site already ranks for it.
Drop anything with zero volume or an intent the site cannot serve.
Cap the output at 50 keywords.
End with the single keyword to build the pillar page around.

---
name: cluster-builder
description: Turn a messy keyword export into clean topic clusters. Use when handed a CSV of hundreds of keywords.
---

Read the export. Cluster by SERP overlap where available, otherwise by shared intent and entity — never by string similarity alone.
One cluster = one page. If two clusters would produce near-identical pages, merge them.
For each cluster: cluster name | primary keyword | supporting keywords | total volume | page type to build.
Cap at 15 clusters. Say how many keywords you dropped and why.
End with the one cluster with the best volume-to-difficulty ratio.

---
name: cannibalisation-finder
description: Find pages competing against each other for the same query. Use when rankings fluctuate or traffic is flat.
---

Pull GSC query-page pairs for the last 90 days. Flag any query where 2+ URLs each took impressions.
Only report cases where the alternate URL took over 10% of impressions — below that it is noise.
Columns: query | URLs | impressions split | clicks | which URL should win and why.
Recommend one of: merge, differentiate intent, or canonicalise. Never all three.
End with the one query costing the most clicks.

---
name: keyword-gap
description: Find keywords a competitor ranks for that you do not. Use for competitive research.
---

Pull the competitor's ranking keywords. Remove anything the site already ranks top 20 for, and anything branded to them.
Keep only keywords the site could credibly rank for given its existing topical authority — say so if a keyword is out of reach.
Columns: keyword | their position | volume | intent | the page you would need to build.
Cap at 25.
End with the one keyword to target this month.

---
name: question-harvester
description: Collect the real questions people ask about a topic. Use before writing an FAQ or GEO content.
---

Gather questions from People Also Ask, GSC queries containing question words, and forum and Reddit threads on the topic.
Deduplicate by meaning, not by wording.
For each: question | volume or frequency signal | whether the site already answers it | best format for the answer.
Cap at 20 questions.
End with the one unanswered question worth a dedicated page.

---
name: brief-builder
description: Write a content brief specific enough that the draft is not generic AI mush. Use before writing any page.
---

Analyse the top 10 for the primary keyword. Build a brief containing: target keyword and cluster, search intent, mandatory format, required H2s with the gap each fills, entities and facts that must appear, the angle no competitor has taken, internal links to include, and the CTA.
Every required section must state WHY it is required, citing what ranks.
Ban generic sections: no "Introduction", no "Conclusion", no "What is X" unless the SERP demands it.
End with the one angle that differentiates this page.

---
name: outline-from-serp
description: Build an H2 outline from what ranks plus what nobody covers. Use after the brief.
---

Extract every H2 and H3 from the top 10 results. Cluster them by topic and count frequency.
Sections appearing in 5+ results are table stakes — include them. Sections in 1-2 results are optional.
Then name what NO result covers but the searcher needs. That is the differentiator.
Output the outline with a one-line note per section on what it must contain.
End with the one section that will make this page better than the current number one.

---
name: first-draft
description: Write the first draft from an approved brief, in the site's voice. Use only after brief-builder.
---

Read the brief and follow it exactly — do not add sections it did not ask for.
Open with the answer, not with context-setting. No "In today's digital landscape".
Every claim gets a specific number, example, or source. If you cannot source it, cut it.
Match sentence length to the site's existing content. Short sentences. Plain words.
End the draft with the CTA from the brief and nothing after it.

---
name: eeat-injector
description: Add real experience and expertise signals to a draft. Use before publishing anything YMYL or competitive.
---

Read the draft. Find every claim stated without evidence.
For each, specify what would make it credible: a named source, a first-hand result, a screenshot, a date, a methodology note, or an author credential.
Columns: claim | current evidence | what to add | where to get it.
Do not invent experience the author does not have — mark those claims for cutting instead.
End with the one addition that most raises the page's credibility.

---
name: content-refresh
description: Build a rewrite plan for a page that is losing traffic. Use when a page decays.
---

Pull the page's GSC data for the last 90 days vs the prior 90. Identify which queries lost position and which are new.
Compare the page to what currently ranks top 3 — what do they cover that it does not, and what has changed since publication.
Output: sections to add, sections to cut, facts to update, title and meta to rewrite.
If the page lost to a different content format entirely, say so — a refresh will not save it.
End with the one change to make first.

---
name: title-rewriter
description: Rewrite titles for pages ranking well with poor CTR. Use monthly.
---

Pull GSC pages in positions 1-10 with CTR below the expected rate for their position.
For each, write 3 title options under 60 characters that carry the primary query, a reason to click, and no clickbait.
Columns: URL | position | current CTR | expected CTR | current title | 3 options.
Cap at 10 pages, sorted by impressions.
End with the one title to ship today.

---
name: meta-description-batch
description: Write meta descriptions for pages that lack one or have a weak one. Use in batches.
---

For each URL, write a meta description of 140-160 characters containing the primary query, the specific benefit, and a verb.
Do not restate the title. Do not use "Learn more about".
Skip any page where Google is already rewriting the description to something better — say which those are.
Cap at 20 URLs.
End with the one page where a better description would earn the most clicks.

---
name: heading-structure-fix
description: Fix broken heading hierarchy that confuses crawlers and readers. Use during on-page audits.
---

For each page, extract the heading tree. Flag: multiple H1s, skipped levels, headings used for styling, and H2s that do not describe their section.
Columns: URL | issue | current heading | corrected heading.
Rewrite headings to carry the query variant they should target, without keyword stuffing.
Cap at 15 pages.
End with the one page whose structure is costing it the most.

---
name: schema-picker
description: Choose the right schema type and generate valid JSON-LD. Use when adding structured data.
---

Identify the page type, then select the schema type that actually earns a rich result for it — not every type does.
Generate valid JSON-LD using only fields the page genuinely contains. Never fabricate ratings, prices, or dates.
State which rich result this makes the page eligible for, and that eligibility is not a guarantee.
Validate required vs recommended properties and flag anything missing.
End with the one page to add schema to first.

---
name: image-alt-batch
description: Write alt text for images missing it, for accessibility and image search. Use in batches.
---

For each image, write alt text under 125 characters describing what the image shows in context.
Decorative images get empty alt, not a description — identify those.
Include the target keyword only where the image genuinely depicts it.
Columns: image | page | current alt | new alt.
End with the one page whose images could realistically earn image-search traffic.

---
name: crawl-blockers
description: Find pages blocked from indexing by robots.txt, noindex, or canonical conflicts. Use for a technical audit or when pages are missing from Google.
---

Pull every indexable URL from the sitemap. For each, check: robots.txt disallow, meta robots noindex, canonical pointing elsewhere, and GSC coverage state.
Report ONLY conflicts — a URL that is submitted but blocked, or canonicalised to a page that is itself noindexed.
Columns: URL | blocker | where it is set | traffic at stake (GSC clicks, last 28d).
Cap at 20 rows, sorted by clicks lost.
End with the single blocker fix that unlocks the most traffic.

---
name: core-web-vitals-triage
description: Rank Core Web Vitals failures by the traffic they actually cost. Use when CWV is failing or a speed audit is requested.
---

Pull CWV field data by URL group. Join each group to its GSC clicks over the last 28 days.
Ignore any failing group under 1% of total clicks — it is noise.
For each remaining group: metric failing | current value | threshold | clicks affected | the one technical cause.
Do not list generic advice like "optimise images" without naming the specific asset or script.
End with the single template to fix first.

---
name: index-coverage-diff
description: Compare GSC index coverage week over week and explain what changed. Use for weekly technical monitoring.
---

Pull GSC index coverage counts by state for this week and the previous week.
Report only states that moved more than 5% or 50 URLs, whichever is larger.
For each: state | last week | this week | delta | most likely cause given what else changed.
Flag any move into 'Crawled - currently not indexed' as a quality signal, not a technical one.
End with the one state to investigate.

---
name: redirect-chain-map
description: Find redirect chains and loops wasting crawl budget and link equity. Use during a migration or technical audit.
---

Crawl the URL list and follow every 3xx to its final destination.
Report only chains of 2+ hops, and any loop.
Columns: start URL | hops | final status | inbound internal links | inbound external links.
Sort by external links, because that is where equity is leaking.
End with the one chain to flatten first.

---
name: sitemap-vs-index
description: Find sitemap URLs Google has not indexed, and say why. Use when new pages are not appearing in search.
---

Diff the sitemap URL set against GSC indexed URLs.
For each missing URL, classify the reason: discovered-not-crawled, crawled-not-indexed, duplicate, soft 404, or blocked.
Group by reason and count. Show 3 example URLs per reason, not all of them.
Crawled-not-indexed means thin or duplicate content — say so plainly rather than suggesting a resubmit.
End with the one reason group to fix.

---
name: orphan-finder
description: Find pages with no internal links pointing to them. Use during a site audit.
---

Build the internal link graph from a crawl. Flag every indexable URL with zero inbound internal links.
Join each orphan to its GSC impressions — an orphan with impressions is a real loss, one with none may deserve to be cut.
Columns: orphan URL | impressions | topic | 3 existing pages that should link to it.
Cap at 20.
End with the one orphan to link first.

---
name: link-opportunity-finder
description: Find existing pages that should link to your money page. Use to push a specific URL.
---

Take the target URL. Find every page on the site that mentions its topic or ranks for a related query but does not link to it.
For each, quote the exact sentence where the link belongs and propose the anchor text.
Columns: source page | its authority signal | sentence to link from | proposed anchor.
Cap at 15 opportunities, sorted by the source page's own traffic.
End with the single highest-value link to add.

---
name: anchor-text-audit
description: Audit internal anchor text for over-optimisation and vagueness. Use during a link audit.
---

Pull every internal anchor pointing to each target URL. Group by anchor text.
Flag two problems: generic anchors ("click here", "read more") and exact-match repetition above 40% of a page's inbound anchors.
Columns: target URL | anchor | count | share | replacement anchor.
End with the one target URL whose anchor profile needs fixing most.

---
name: hub-spoke-builder
description: Turn a topic cluster into a linked hub-and-spoke structure. Use after cluster-builder.
---

Take the cluster. Designate one pillar page and assign every other page as a spoke.
Specify: pillar links down to every spoke, every spoke links up to the pillar, and spokes link sideways only where topically adjacent.
Output the full link map with the anchor text for each link.
Flag any spoke that duplicates the pillar's intent — that one should be merged, not linked.
End with the one link that completes the structure.

---
name: link-depth-fix
description: Find pages buried more than 3 clicks from the homepage. Use when deep pages underperform.
---

Crawl from the homepage and record click depth for every indexable URL.
Flag pages at depth 4+ that have GSC impressions — those are earning attention despite being buried.
Columns: URL | depth | impressions | the shallow page that should link to it.
Do not recommend adding links to the main navigation for every page — name the one or two that deserve it.
End with the one page to surface.

---
name: link-gap
description: Find sites linking to competitors but not to you. Use for link prospecting.
---

Pull backlink profiles for 3 competitors. Keep only referring domains linking to 2+ of them but not to you.
Remove directories, scrapers, and anything with no real traffic.
Columns: domain | authority | which competitors it links to | the page it linked to | why it would link to you.
Cap at 25 prospects.
End with the one prospect most likely to say yes.

---
name: broken-link-prospector
description: Find broken pages on relevant sites you could replace with your content. Use for outreach.
---

Find pages on target sites that link out to URLs returning 404.
Keep only broken targets whose topic matches content you already have or could write in a day.
Columns: host page | broken URL | its former topic | your replacement URL | inbound links the broken URL still has.
End with the one replacement worth pitching first.

---
name: digital-pr-angle
description: Turn your own data into a story a journalist would actually cover. Use for link earning.
---

Take the available proprietary data. Find the angle that is genuinely newsworthy: a counterintuitive finding, a ranking or league table, a trend with a number attached, or a seasonal hook.
State the headline a journalist would write, the single statistic that carries it, and the 3 publications that cover this beat.
If the data does not support a real story, say so — do not manufacture one.
End with the one angle to pitch.

---
name: outreach-writer
description: Write outreach emails that get replies. Use after any prospecting skill.
---

For each prospect, write an email under 120 words. Reference the specific page and what is actually on it, state the value in one sentence, make one ask.
No compliments about their "amazing blog". No "I hope this email finds you well". No follow-up threat.
Subject line under 45 characters, lowercase, specific.
End with the one prospect to send to first.

---
name: toxic-link-triage
description: Decide whether a suspicious backlink profile needs action. Use when links look spammy.
---

Classify referring domains as: natural, low quality but harmless, or genuinely manipulative.
Google ignores most spam automatically — recommend disavow only for links that are paid, part of a scheme, or arrived alongside a manual action.
Columns: domain | classification | evidence | action.
If there is no manual action and no pattern of paid links, recommend doing nothing and say why.
End with the one decision: disavow or leave it.

---
name: position-8-15-finder
description: Find pages sitting at position 8-15 that deserve a push. Use first — this is the fastest win on the list.
---

Pull GSC queries from the last 28 days with average position between 8 and 15 and impressions above 100.
Rank by impressions multiplied by the CTR you would gain moving to position 3.
Columns: URL | query | position | impressions | estimated clicks gained | the one change to make.
The change must be specific: a section to add, a link to build, a title to rewrite. Not "improve the content".
Cap at 10 rows. End with the single page to push first.

---
name: low-ctr-finder
description: Find pages ranking in the top 10 with a CTR far below expectation. Use monthly.
---

Pull GSC pages in positions 1-10. Compare each page's CTR to the benchmark for its position.
Flag only pages more than 40% below benchmark with over 200 impressions.
Columns: URL | position | actual CTR | expected CTR | clicks left on the table | likely cause.
A position-4 page with 1.2% CTR is a packaging problem, not a ranking problem — say which it is.
End with the one page to rewrite.

---
name: decay-detector
description: Find pages losing traffic before the decline becomes serious. Use weekly.
---

Compare the last 28 days to the prior 28, and to the same period last year to control for seasonality.
Flag pages down more than 20% that are not explained by seasonality.
For each: URL | clicks then | clicks now | position change | whether it lost position or lost impressions.
Losing impressions at a stable position means the SERP changed, not your page — distinguish the two.
End with the one page to refresh.

---
name: query-to-page-map
description: Map which page Google actually serves for each of your target queries. Use to catch mismatches.
---

For each target query, pull the URL GSC reports as ranking.
Compare it to the URL you intended to rank. Flag every mismatch.
Columns: query | intended URL | actual URL | position | why Google chose differently.
A mismatch usually means intent or internal linking, not keyword density.
End with the one query where the wrong page is ranking and it matters most.

---
name: monthly-report
description: Produce a monthly SEO report a client will actually read. Use at month end.
---

Pull clicks, impressions, average position and indexed pages for the month and the prior month.
Structure: what moved and by how much, what we shipped, what it produced, what is next.
Every number carries its absolute value alongside the percentage. Every win names the action that caused it.
No screenshots of dashboards. No metric without a decision attached to it.
Cap at one page. End with the one priority for next month.

---
name: intent-classifier
description: Classify keywords by search intent so you build the right page type. Use before any content decision.
---

For each keyword, classify as informational, commercial investigation, transactional, or navigational.
Base the call on what actually ranks in the SERP, not on the wording of the keyword.
Columns: keyword | intent | evidence from the SERP | correct page type.
Flag any keyword where the SERP shows mixed intent — those need a hybrid page or should be skipped.
End with the one keyword whose intent the site is currently getting wrong.

---
name: serp-shape-reader
description: Read what content format actually ranks for a keyword. Use before writing anything.
---

Analyse the top 10 for the keyword. Record: dominant format (listicle, guide, tool, product, video, forum), median word count, and how many results are from the same site type.
If 6+ results share one format, that format is mandatory — say so.
Note which SERP features occupy space above the fold.
Do not summarise the content of each result; report the pattern.
End with the one format decision for this page.

---
name: intent-mismatch
description: Find pages ranking for queries they were not built to serve. Use when a page has impressions but no clicks.
---

Pull GSC queries per page. For each page, compare the intent of its top queries to the page's actual type.
Flag pages where the dominant query intent does not match the page type — a product page pulling informational queries, or a blog post pulling transactional ones.
Columns: URL | page type | dominant query intent | impressions | clicks | CTR.
Cap at 10 pages.
End with the one page to repurpose.

---
name: buying-stage-mapper
description: Map keywords to funnel stage so content matches where the buyer is. Use when planning a content calendar.
---

For each keyword, assign a stage: unaware, problem-aware, solution-aware, product-aware, or ready to buy.
For each: keyword | stage | what the reader needs at that moment | the CTA that fits.
Flag any stage with no coverage on the site — that gap is the finding.
Do not assign a hard-sell CTA to a problem-aware keyword.
End with the one funnel stage to fill first.

---
name: serp-feature-audit
description: Find SERP features you could realistically win. Use to capture clicks without moving position.
---

For each target keyword, list which SERP features appear: featured snippet, PAA, image pack, video, local pack, AI Overview.
For each feature the site does not hold, judge whether it is winnable given current position — anything outside the top 10 is not.
Columns: keyword | feature | current holder | your position | the specific change needed to win it.
Cap at 15.
End with the one feature to go after.

---
name: gbp-audit
description: Audit a Google Business Profile for completeness and ranking signals. Use for any local business.
---

Check the profile for: primary and secondary categories, full description, hours including holiday hours, service area, attributes, products or services listed, photo count and recency, Q&A, and post frequency.
Columns: element | current state | impact on local ranking | fix.
Category choice outranks almost everything else — check it first and say if it is wrong.
End with the one element to fix today.

---
name: nap-consistency
description: Find name, address and phone inconsistencies across the web. Use before any local campaign.
---

Collect the business NAP as listed on the site, GBP, and the major directories and aggregators.
Compare against the canonical version character by character — abbreviations, suite numbers, and phone formats all count.
Columns: source | name | address | phone | mismatch | how to correct it.
End with the one citation to correct first, weighted by the directory's authority.

---
name: local-landing-page-builder
description: Build a location page that is not thin duplicate content. Use when targeting multiple cities.
---

For the target location, produce a page outline containing: locally specific service detail, named neighbourhoods or districts served, a local case study or customer, location-specific FAQs, embedded map, and NAP.
At least 40% of the body must be content that could not appear on any other location page. Name exactly which sections carry that.
If you cannot find enough local specifics, say the page should not be built.
End with the one local detail that makes this page unique.

---
name: review-response-writer
description: Write responses to customer reviews that help local ranking. Use weekly.
---

For each review, write a response under 60 words. Thank by name, address the specific point raised, and where natural include the service and city once.
Negative reviews: acknowledge, take it offline with a contact route, never argue, never blame the customer.
Do not use the same opening line twice across responses.
End with the one review that needs a reply today.

---
name: local-pack-gap
description: Work out why a competitor outranks you in the map pack. Use when local rankings stall.
---

Compare your GBP to the top 3 map pack competitors on: primary category, review count, average rating, review recency, photo count, post frequency, proximity to the searcher, and website relevance for the query.
Columns: signal | you | competitor 1 | competitor 2 | competitor 3 | gap.
Proximity cannot be changed — if that is the whole gap, say so honestly.
End with the one closable gap.

---
name: citation-checker
description: Check whether AI engines cite your brand for your key questions. Use monthly.
---

Take the 10 questions your buyers actually ask. Query each against ChatGPT, Claude and Perplexity.
Record: are you cited, which competitors are, which source URL the engine used, and how your brand is described.
Columns: question | cited yes/no | competitors cited | source used.
Do not treat one absent citation as a trend — note the sample size.
End with the one question where a citation is most winnable.

---
name: competitor-citation-teardown
description: Work out why an AI engine cites a competitor instead of you. Use after citation-checker.
---

Take a question where a competitor is cited. Fetch the exact page the engine used.
Compare it to your equivalent page on: directness of the answer in the first 100 words, presence of a quotable standalone passage, structured data, freshness date, named author, and third-party corroboration elsewhere on the web.
Columns: signal | their page | your page | gap.
End with the one change that would make your page the better citation.

---
name: answer-block-writer
description: Write passages an AI engine can lift and cite. Use on every page you want cited.
---

For each target question, write a 40-60 word answer block that stands alone with no surrounding context, states the answer in the first sentence, includes one specific number or named entity, and reads as a complete quote.
Place it directly under a heading phrased as the question.
No hedging, no "it depends" openers, no marketing language.
End with the one page to add an answer block to first.

---
name: entity-consistency
description: Check that AI engines describe your brand accurately and consistently. Use quarterly.
---

Ask each engine to describe the brand, its category, its pricing, and its main alternatives.
Compare answers against the truth. Flag every inaccuracy and where the engine likely picked it up.
Columns: claim made | accurate? | likely source | correction route.
Corrections come from fixing the source, not from prompting the engine — say which source to fix.
End with the one inaccuracy to correct first.

---
name: llms-txt-builder
description: Generate an llms.txt that tells AI crawlers what matters on your site. Use once, then maintain.
---

Build llms.txt listing the site's purpose in one line, then the pages that genuinely represent it: docs, pricing, key guides, and about.
One line per URL with a short description of what it answers. Exclude blog archives, tags, and thin pages.
Keep it under 50 URLs — this is a curation, not a sitemap.
Note that adoption is not universal and this is a low-cost bet, not a ranking factor.
End with the one page that most deserves to be in it.

You are helping me expand a seed keyword list before I run it through a keyword tool. Below are my 3 to 10 seed keywords, one line describing what the site sells, and one line on who buys it. If any of the three is missing, ask me for it before doing anything.

For each seed, generate candidates in four buckets: modifiers (best, vs, alternative, pricing, template, example), questions (how, why, what, can), audience (for small business, for agencies, for beginners), and job to be done (the verb a buyer would type). Skip any candidate that is only a plural or a word order swap of another. Skip candidates that describe a different product category than the one I described.

Output one table, max 40 rows, columns: candidate, seed, bucket, likely intent (informational, commercial, transactional), why a buyer types it (max 10 words). Do not write search volumes or difficulty scores. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single bucket I should paste into my keyword tool first and why, no alternatives.

Below is my keyword export (Ahrefs or Semrush) with columns keyword, volume, difficulty, and the URL currently ranking for me if any. If the export is not pasted below, ask me for it before doing anything.

Cluster the keywords so that one page can rank for every keyword in a cluster. Two keywords belong together when a searcher would be satisfied by the same page: same intent, same entity, same format. Treat modifiers like best, vs, pricing and how to as different intents and keep them in separate clusters even when the head term matches. Move keywords with volume under 10 to an ignored list. Rank clusters by summed volume, then by the lowest median difficulty.

Output a table, max 25 clusters, columns: cluster name, primary keyword, keyword count, summed volume, median difficulty, current ranking URL if one exists, page type (guide, comparison, product, tool). Below it, give the ignored keywords as a count only. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one cluster to build or optimise first, no alternatives.

Below is my Search Console export for the last 3 months with one row per query and page pair, columns query, page, clicks, impressions, CTR, position, plus my brand terms. If the export or the brand terms are missing, ask me to paste them before doing anything.

Find cannibalisation: queries where two or more of my pages each have at least 50 impressions. Exclude brand queries. Exclude queries under 100 total impressions. For each query, name the page with the most clicks as the leader and the others as competitors. Mark the case serious when the leader holds under 60 percent of the query's clicks or when the pages sit within 3 positions of each other.

Output a table, max 15 rows, sorted by total impressions descending, columns: query, total impressions, leader page, competitor pages, leader click share, severity (serious or minor), fix (consolidate, redirect, differentiate intent, or canonical). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single query to fix first and which fix, no alternatives.

Below is a keyword gap export (Ahrefs Content Gap or Semrush Keyword Gap) listing keywords where at least one competitor ranks and I do not, with columns keyword, volume, difficulty, and each competitor's position. If it is missing, ask me for it before doing anything.

Keep only keywords where at least two competitors rank in the top 10, volume is at least 50, and the keyword is not a competitor brand or product name. Group what remains by topic. For each group say whether I already have a page that could rank; ask for my page list if I have not pasted one. Score each group by competitors covering it, summed volume and median difficulty.

Output a table, max 20 rows, one per topic, sorted by summed volume descending, columns: topic, example keyword, keyword count, summed volume, median difficulty, competitors in top 10, my existing page or "none". If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one topic to create or expand a page for first, no alternatives.

Below are two inputs: my Search Console Queries export for the last 12 months (query, clicks, impressions, position) and a list of People Also Ask questions I collected for my main topic. If either is missing, ask me for it before doing anything.

Harvest questions. From the export, keep only queries that start with how, what, why, when, which, can, does, is, or should, and have at least 30 impressions. From the People Also Ask list, keep every question. Merge duplicates worded differently. Mark each question as answered (I rank in position 1 to 10), weak (position 11 to 30), or missing (no position, or only in the People Also Ask list).

Output a table, max 30 rows, sorted by impressions descending with People Also Ask only rows at the end, columns: question, source (GSC, PAA, both), impressions, position, status, suggested home (existing page URL or new page). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one question to answer this week and where, no alternatives.

Below are: my target keyword, the titles and H2 headings of the top 10 ranking pages copied from the SERP, and who the reader is. If any of these is missing, ask me for it before doing anything.

Write a brief for a writer who does not know SEO. Include: the search intent in one sentence; the angle that separates this page from the 10 pasted (something at most 2 of them cover); a target length as a range taken from the pasted pages only, and if I did not give their lengths write "cannot compute from this data"; a heading outline with H2s and H3s, max 8 H2s; entities and terms to include, max 15; questions to answer, max 6; internal pages to link (ask for my sitemap if needed); what to avoid, max 4 items taken from the top 10's weaknesses.

Use exactly these sections in this order: Intent, Angle, Length, Outline, Entities, Questions, Internal links, Avoid. Never estimate a number that is not in the data.

Finish with one action: the single section the writer should draft first, no alternatives.

Below are the full heading structures (H1 to H3) of the top 5 pages ranking for my keyword, plus the keyword itself. If I have pasted fewer than 3 pages or no keyword, ask me for the rest before doing anything.

Build one outline that beats them. List every heading topic that appears in at least 3 of the 5 pages: these are mandatory. List topics that appear in only 1 or 2 pages: these are optional, keep the ones a reader with this intent needs. Propose max 3 topics none of the 5 cover, from what a reader asks next. Put the answer to the query before the background. Cap the outline at 10 H2s, each with 0 to 3 H3s.

Output three fixed sections: Coverage table (topic, pages covering it out of 5, mandatory or optional), Proposed outline, Gaps (max 3, one line each on why it earns a citation). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one section to write first, no alternatives.

Below are: a content brief with its outline, and two published articles from my site for voice. If the brief or the samples are missing, ask me for them before doing anything.

Write the first draft. Follow the outline exactly: do not add or drop H2s. Match the samples on sentence length, person (I, we, or you), and how much they hedge. The first paragraph answers the query in under 60 words with no preamble. Each H2 section opens with its direct answer, then supports it. Use only facts that are in the brief or the samples. Where the brief needs a statistic not given, write the placeholder [NEED SOURCE: what is needed] instead of a number. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate. No introduction over 3 sentences, no section headed Conclusion, no bullet list over 5 items.

Output: the draft as markdown, then a table of max 10 rows listing every placeholder with the section it sits in.

Finish with one action: the single placeholder I should source first, no alternatives.

Below is a draft article and a short list of facts about the author and the company: what we have done ourselves, data we own, credentials, and customers we can name. If the draft or the facts list is missing, ask me for it before doing anything.

Find where the draft makes a claim any generic writer could make and replace it with a signal only we can give. Signal types: a first hand result with a number we own, a screenshot or dataset we could add, a named customer example, a limitation we hit ourselves, a quote from a person on the facts list. Do not invent experiences, numbers, names or credentials that are not on the list. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output a table, max 10 rows, ordered by how much the change strengthens the page, columns: section heading, current sentence (quoted), replacement sentence, signal type, fact used. Then an author box draft, max 60 words, using only listed credentials.

Finish with one action: the one edit to make first, no alternatives.

Below are: the current text of a page that lost traffic, and its Search Console Queries export for the last 3 months compared with the previous 3 months (query, clicks, impressions, position for both periods). If either is missing, ask me for it before doing anything.

Diagnose first, then plan. Split the loss three ways: queries where position dropped by 3 or more, queries where position held but impressions fell by 30 percent or more (demand or SERP change, not the page), and queries that disappeared. Ignore queries under 20 impressions in both periods. Read the page against the dropped queries: what does it fail to answer, what is stale (dates, prices, versions, screenshots)?

Output: a Diagnosis section (max 5 lines), a table of max 10 queries sorted by clicks lost, columns: query, clicks before, clicks after, position before, position after, cause, then a Refresh plan of max 6 edits naming the section and the change. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single edit to make first, no alternatives.

Below is my Search Console Pages export for the last 28 days (page, clicks, impressions, CTR, position) with the current title tag as a column, plus my brand name. If either is missing, ask me to paste it before doing anything.

Find pages where CTR falls below these floors: position 1 to 3, 15 percent; 4 to 6, 6 percent; 7 to 10, 3 percent. Keep pages with at least 200 impressions and position 10 or better. Exclude pages whose top query is branded. Rank by impressions multiplied by the CTR gap.

For each, write one new title of 50 to 60 characters: lead with the query the page already ranks for, state the format (guide, template, checklist, comparison), drop the brand name except on the home page. No clickbait, no year unless updated yearly.

Output a table, max 10 rows, columns: page, impressions, position, current CTR, expected CTR, current title, new title, character count. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single title to change today, no alternatives.

Below is a list of up to 20 URLs, each with its H1, its current meta description (blank if none) and its top Search Console query. If either is missing, ask me for it before doing anything.

Write one meta description per URL, 120 to 155 characters, that starts with the outcome the reader gets, contains the top query in natural wording within the first 100 characters, matches the page type (a product page says what it is and for whom, a guide what it teaches, a comparison what is compared), and ends with a concrete reason to click, not "learn more". Do not repeat the H1 verbatim. Do not claim numbers, prices or awards absent from the input. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Flag URLs whose current description is blank, duplicated, or under 70 characters, and put those first.

Output a table, max 20 rows, columns: URL, flag reason, new description, character count, top query included (yes or no).

Finish with one action: the one URL to update first, no alternatives.

Below is the heading list of one page in order, each line as heading level then heading text, plus the page's target keyword. If I have pasted only text with no levels, or no keyword, ask me for it before doing anything.

Audit the hierarchy. Check: exactly one H1 and it contains the target keyword or a clear synonym; no level is skipped (H2 to H4 with no H3 between); no heading is a styling choice (short bold labels like Note or Tip set as H3); headings are not questions where a plain statement is clearer, unless the section answers a People Also Ask question; every H2 states its subject so someone reading only headings understands the page; no two headings say the same thing.

Output a table, max 15 rows, one per problem, columns: current heading, level, problem, proposed heading, proposed level. Then the full corrected outline. Do not add sections the page does not have. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single heading change with the most effect, no alternatives.

Below are: the URL and full visible text of one page, and its page type if I know it (article, product, FAQ, how to, local business, software). If the text is missing, ask me to paste it before doing anything.

Pick the schema types this page can honestly carry and draft the JSON-LD. Rules: choose only types whose required properties are visible on the page (a FAQPage needs visible question and answer pairs, a Product with offers needs a visible price, a HowTo needs numbered steps). Do not add ratings, review counts, prices, dates or author names that are not in the text. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate. Prefer one main type plus at most two supporting types.

Output: a table with max 3 rows, columns: schema type, why it fits, required properties found, properties missing; then one JSON-LD block using a single @graph; then a list of max 5 things to verify before deploying.

Finish with one action: the single property I must confirm on the page before adding this markup, no alternatives.

Below is a list of images: each row has the filename, the page URL, the nearest heading, and the current alt text (blank if none). If the list is missing page context, ask me for it before doing anything.

Write one alt text per image, 4 to 16 words, describing the image for someone who cannot see it, naming the subject specifically (the product, the chart, the screen), and using the page's topic term only where the image actually shows it. Rules: no "image of" or "picture of", no keyword lists, decorative images (dividers, icons) get an empty alt marked decorative, charts state their finding, screenshots name the tool and the screen. Do not guess what an image shows: if the filename and context do not tell you, write "needs a look". If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output a table, max 30 rows, columns: image, page, current alt, new alt, type (content, chart, screenshot, decorative, needs a look).

Finish with one action: the one image to fix first, no alternatives.

Below are: my robots.txt, and a Screaming Frog internal HTML export with columns Address, Status Code, Indexability, Indexability Status, Meta Robots, Canonical Link Element, and Inlinks. If the export or the robots file is missing, ask me for it before doing anything.

Find what stops important pages from being indexed. Check, in this order: URLs blocked by a robots rule that also carry a noindex (Google cannot see the noindex); noindexed URLs with 5 or more inlinks; canonical conflicts (the canonical target is itself noindexed, redirected, non 200, or canonicalised elsewhere); indexable URLs with 0 inlinks; canonicals pointing elsewhere on pages that receive traffic, and ask for my Search Console pages if you need them. Ignore URLs with parameters unless they have inlinks.

Output a table, max 20 rows, ordered by inlinks descending, columns: URL, problem type, evidence (the exact directive or canonical target), inlinks, fix. Then a count per problem type. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single blocker to remove first, no alternatives.

Below are: 75th percentile field data (LCP, INP, CLS from PageSpeed Insights or the Search Console Core Web Vitals report) per URL or URL group, and sessions or clicks per URL from GA4 or Search Console for the last 28 days. If the traffic numbers are missing, ask me for them before doing anything.

Triage by traffic, not by score. A URL fails when LCP is above 2.5 seconds, INP above 200 milliseconds, or CLS above 0.1. Group failing URLs by template (product, category, article, home) when paths make it obvious. Rank groups by traffic on failing URLs. For each group say which metric fails and which is within 20 percent of its threshold, the cheapest win. Name causes only if I pasted lab diagnostics; otherwise write "needs lab data".

Output a table, max 10 rows, columns: template or URL, URLs failing, traffic on failing URLs, failing metric, value, distance to threshold, likely cause. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single template and metric to fix first, no alternatives.

Below are two exports of the Search Console Pages indexing report (URL and its status or reason), one from last week and one from this week. If I pasted only one, ask me for the other before doing anything.

Diff them. Find: URLs that moved from indexed to not indexed, with the new reason; URLs that moved from not indexed to indexed; URLs new this week; URLs that disappeared from the report. Group by reason (crawled currently not indexed, discovered currently not indexed, noindex, redirect, soft 404, duplicate with Google chosen canonical, not found). Ignore groups under 5 URLs unless a URL sits in a money path, and ask for my key paths if I have not listed them. Rank reasons by URLs lost.

Output a table, max 10 rows, columns: reason, URLs lost this week, URLs gained, example URL, likely trigger (deploy, template change, robots change, or "unknown, needs a look"). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single reason group to investigate first, no alternatives.

Below is a Screaming Frog redirect chains report (each row: a starting URL, every hop with its status code, the final URL). If it is missing, ask me for it before doing anything.

Map the chains. Flag: chains of 2 or more hops; loops; chains ending in a 4xx or 5xx; chains mixing 301 and 302; http to https and www hops done separately (fold into one). For each start URL, state the final destination and the one direct redirect that replaces the chain. If I also paste an inlinks export, count internal links pointing at the first hop; those should be updated to the final URL.

Output a table, max 20 rows, sorted by hop count descending then by inlinks, columns: start URL, hops, status codes in order, final URL, final status, fix (collapse to one 301, fix loop, fix broken target, update internal links). Then totals: chains found, loops, broken endings. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single chain to collapse first, no alternatives.

Below are: the list of URLs in my XML sitemap, and a Screaming Frog crawl export (Address, Status Code, Indexability, Canonical Link Element) or the Search Console Pages report export. If either is missing, ask me for it before doing anything.

Compare the two sets. Report: sitemap URLs that return non 200, are noindexed, or canonicalise elsewhere (they should leave the sitemap); indexable 200 URLs found in the crawl but absent from the sitemap (they should join it, unless they are parameter or paginated URLs); sitemap URLs never seen in the crawl (orphans, check whether they should exist). Ignore image, PDF and feed URLs. Where the Search Console export is present, add how many sitemap URLs are reported as not indexed and the top reason.

Output three tables, max 15 rows each, columns: URL, issue, evidence (status, directive or canonical target), fix. Then a totals line: sitemap URLs, should leave, should join, orphans. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one set of URLs to fix in the sitemap first, no alternatives.

Below are: a Screaming Frog crawl export with Address, Status Code, Indexability, Inlinks and Crawl Depth; my XML sitemap URL list; and, if I have it, a Search Console Pages export for the last 3 months. If the crawl or the sitemap is missing, ask me for it before doing anything.

Find orphans and near orphans. Orphan: an indexable 200 URL in the sitemap or in Search Console with 0 inlinks in the crawl. Near orphan: 1 to 2 inlinks, or inlinks only from paginated, tag or footer pages. Exclude legal pages, author pages and URLs with parameters. Rank by impressions if given, otherwise by whether the URL sits in a money path (ask for my key paths if needed).

Output a table, max 20 rows, columns: URL, inlinks, crawl depth, impressions, type (orphan, near orphan), suggested linking page (a topically close crawl page with at least 5 inlinks), proposed anchor. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single orphan to link first and from where, no alternatives.

Below are: the URL and target keyword of one money page, and a list of my other pages with URL, title, top 3 Search Console queries and impressions. If the money page or the page list is missing, ask me for it before doing anything.

Find internal link opportunities. A candidate qualifies when its title or queries share the money page's topic and it does not already link there. Ask for the money page's inlinks; if I cannot give them, mark that column unknown. Prefer candidates with more impressions; they are crawled and read more. For each candidate write the anchor (3 to 7 words, containing the target keyword or a close variant, no two identical) and the body sentence it should live in, never footer or sidebar.

Output a table, max 10 rows, ordered by impressions descending, columns: candidate URL, shared topic, impressions, already links (yes, no, unknown), anchor, sentence. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single link to add first, no alternatives.

Below is an internal links export (Screaming Frog All Inlinks or similar) with columns Source, Destination, Anchor, and Link Position if available. If it is missing, ask me for it before doing anything.

Audit anchors on the most linked destinations. For each destination with at least 5 inlinks, count its anchors. Flag: generic anchors (click here, read more, here, this page, learn more) as a share of the page's inlinks; one exact keyword anchor on more than 60 percent of inlinks (over optimised); anchors describing a different topic from the destination's title; the same anchor pointing to two different destinations. If Link Position is present, keep body links only and skip navigation and footer.

Output a table, max 15 rows, sorted by inlinks descending, columns: destination, inlinks, generic share, top anchor and its share, distinct anchors, flags, proposed anchor variants (max 3, each 3 to 7 words). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one destination whose anchors to rewrite first, no alternatives.

Below are my pages on one topic (URL, title, top Search Console query, impressions), plus the topic name. If it has under 4 pages or no impressions, ask me for a fuller list first.

Choose the hub: the page whose query is the broadest head term with the most impressions, or say a new hub is needed if none targets it. Every other page is a spoke. Rules: each spoke links up to the hub with an anchor containing the head term; the hub links down to every spoke, saying what each answers; spokes link sideways only to the 2 spokes closest in sequence; off topic pages are excluded and listed. Merge spokes whose queries one page could answer.

Output: a Hub line (URL or "new page needed", plus the head term), a table of max 15 spokes, columns: spoke URL, sub topic, anchor to hub, sideways links (max 2), merge into (or none), then an Excluded list. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single link to add first, no alternatives.

Below is a Screaming Frog crawl export (Address, Crawl Depth, Inlinks, Indexability) and, if I have it, a Search Console Pages export with clicks and impressions. If the crawl is missing, ask me for it before doing anything.

Too deep means crawl depth 4 or more (home page is 0). Rank by impressions if given, otherwise by inlinks. Ignore paginated URLs, parameters and legal pages. For each, propose the shallowest existing page that could link to it (a category, a hub, a depth 1 to 2 page on the same topic); the new depth is that page's depth plus 1. If over 30 percent of indexable URLs sit at depth 4 or more, say so and name the structural cause (pagination only, missing category links, no related links).

Output a table, max 20 rows, columns: URL, current depth, inlinks, impressions, proposed linking page, new depth. Then a depth distribution table (depth, URL count). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single page to surface first and from where, no alternatives.

Below is a referring domains export for my site and for 2 to 4 competitors (Ahrefs Link Intersect, Semrush Backlink Gap, or one export per site), with columns domain, authority score, the linking page where given, and which sites it links to. If the exports are missing or cover only my site, ask me for the competitor data before doing anything.

Find domains that link to at least 2 competitors and not to me. Exclude directories, forums, coupon sites, domains with authority under 20, and properties owned by a competitor. Classify each remaining domain by how it probably linked (editorial mention, resource list, guest post, tool roundup, partner page), read from the linking page URL when given. Rank by the number of competitors linked, then by authority.

Output a table, max 20 rows, columns: domain, authority, competitors linked (count and names), link type, linking page example, angle for me (max 8 words). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single domain to contact first and with which angle, no alternatives.

Below is a list of dead pages (404s or expired domains) in my niche: for each, the URL, what it was about, and an export of pages still linking to it (referring page, domain authority, anchor). If the linking page export is missing, ask me for it before doing anything.

For each dead page, match it to my page on the same subject (ask for my page list if needed). Skip dead pages with no equivalent; do not propose new pages. Keep referring pages with authority 20 or more, drop link lists with more than 50 outbound links, and drop referring pages on the dead site's own domain. Rank by authority.

Output a table, max 20 rows, columns: referring page, authority, dead URL it links to, anchor used, my replacement page, fit (exact, close, weak). Then one draft email, max 90 words, for exact fit rows, naming the broken link and offering the replacement, no flattery. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one referring page to email first, no alternatives.

Below is a dataset I own (a CSV or table from my product, my customers, or my own research) and one line on who I want coverage from. If the dataset is missing, ask me to paste it before doing anything; do not propose angles from public statistics.

Find the story. Compute only what the columns allow: counts, shares, medians, rank orders, and changes between two periods if both are present. Do not extrapolate to a population the data does not cover. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate. An angle qualifies when it is surprising against what the target audience assumes, specific (one number, one comparison), and defensible from the rows.

Output a table, max 5 rows, columns: headline (max 12 words), the one number, how it was computed from the columns, who would cover it (publication type), the weakness a journalist would raise. Then a Methodology paragraph, max 80 words, stating sample size, period, and what was excluded.

Finish with one action: the single angle to pitch and to whom, no alternatives.

Below are: a list of link prospects (site, target page, why my page fits it, contact first name if known), and my page's URL with a one line description. If the "why it fits" column is missing, ask me for it before doing anything.

Write one email per prospect, max 110 words. Structure: subject under 7 words naming their page; first sentence names the specific thing on their page mine adds to (a missing step, an outdated number, a broken reference); second sentence says what my page is in plain words; one sentence on where the link helps their reader; a close asking a yes or no question. No compliments, no "hope this finds you well", no mention of SEO or link exchange. Do not claim numbers about my page absent from the input. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output a table, max 10 rows, columns: prospect, subject, email body, follow up line (one sentence for 5 days later).

Finish with one action: the one prospect to send first, no alternatives.

Below is a backlinks export (Ahrefs, Semrush or Search Console Links) with columns referring page, domain, authority, anchor, first seen, and dofollow or nofollow if present. If it is missing, ask me for it before doing anything.

Triage without over disavowing. Bucket every domain as harmless (nofollow, or authority under 10 with a generic anchor), watch (exact match money anchors from unrelated sites, or 20 or more links from one domain in a short window), or act (link farms, hacked pages, foreign language spam with commercial anchors, paid network patterns like one template across many domains). Volume alone is not toxic. Flag a spike when over 30 percent of first seen dates fall inside 30 days.

Output a table, max 20 rows, watch and act only, act first then by authority, columns: domain, links, authority, anchor sample, first seen, bucket, reason. Then totals per bucket and a domain: format disavow list for the act bucket only. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single step to take with the act bucket, no alternatives.

Below is my Search Console export for the last 3 months with query and page dimensions, columns query, page, clicks, impressions, CTR, position, plus my brand terms. If the export or the brand terms are missing, ask me to paste them before doing anything.

Find striking distance queries: average position between 8 and 15, at least 100 impressions, not a brand query. Rank by the clicks gained if the query moved to position 5, using a CTR of 6 percent at position 5, and state that assumption in the output. Group queries by page so I see each page's total opportunity. Note when a page has 3 or more striking distance queries, since one improvement lifts them all.

Output a table, max 15 rows, one per page, columns: page, striking distance queries (count), top query, impressions total, position of the top query, projected clicks at position 5, what the page is missing versus the query (one line). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single page to work on this week, no alternatives.

Below is my Search Console Queries export for the last 28 days (query, clicks, impressions, CTR, position) and my brand terms. If either is missing, ask me for it before doing anything.

Find queries in the top 10 where CTR sits below the position's expected curve. Use this curve unless I give my own: position 1, 25 percent; 2, 15; 3, 10; 4, 7; 5, 6; 6 to 10, 3. Keep queries with position 10 or better and at least 100 impressions. Exclude brand queries. Exclude queries where the SERP probably answers above the results (the query reads as a definition, a number, a date, a unit conversion) and list those separately with a note. Rank by impressions multiplied by the CTR gap in points.

Output a table, max 15 rows, columns: query, impressions, position, CTR, expected CTR, missed clicks (impressions times gap), page, likely cause (title, snippet, SERP feature, intent mismatch). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single query and page to fix first and which element, no alternatives.

Below is my Search Console Pages export comparing the last 3 months with the previous 3 months (page, clicks, impressions and position for both periods). If I pasted a single period, ask me for the comparison export before doing anything.

Detect decay, not noise. A page decays when clicks fell by 20 percent or more and by at least 30 clicks, and position worsened by 1 or more. Separate from that: pages where clicks fell but position held (demand fell or a SERP feature appeared), and pages where position held, impressions rose and CTR fell (something pushed us down visually). Ignore pages under 100 clicks in the earlier period. Rank by clicks lost.

Output a table, max 15 rows, columns: page, clicks before, clicks after, change in percent, position before, position after, pattern (decay, demand drop, visual push), first check (refresh, SERP, indexing). Then total clicks lost across decaying pages. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single page to open first and what to check, no alternatives.

Below is my Search Console export with both query and page dimensions for the last 3 months (query, page, clicks, impressions, position), plus my brand terms. If the export has only queries or only pages, ask me for the combined one before doing anything.

Build the query to page map. For each query with at least 50 impressions, name the page that should own it: the page with the most clicks when it also holds the best position, otherwise flag a conflict. Then, for each owning page, list its query set and mark queries that do not fit the page's topic (the page ranks by accident). Mark queries with no page at position 20 or better as unowned. Skip brand queries.

Output a table, max 25 rows, columns: query, impressions, owning page, position, status (owned, conflict, misfit, unowned), action (keep, consolidate, add a section, new page). Then a count per status. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single conflict or unowned query to resolve first, no alternatives.

Below are: my Search Console totals for this month, last month and the same month last year (clicks, impressions, CTR, position), this month's top 20 pages and queries with comparison columns, and GA4 organic conversions if I have them. If the comparison columns are missing, ask me for them before doing anything.

Write the report for someone who reads it in 2 minutes. Compare month over month and year over year; say which matters more this month (seasonality shows year over year). Name the 3 pages gaining the most clicks and the 3 losing the most, with position change. Name new queries with at least 200 impressions. Do not explain a change with a cause the data does not show; write "unknown, needs a look".

Output fixed sections: Headline (one sentence, the click change), Numbers table (metric, this month, last month, last year, change), Winners (max 3), Losers (max 3), New queries (max 5), Next month. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action under Next month: the single thing to do, no alternatives.

Below is a list of keywords, one per line, with volume if I have it. If the list is missing, ask me to paste it before doing anything.

Classify each keyword's dominant intent using only the words in the query: informational (how, what, why, guide, examples), commercial investigation (best, top, vs, review, alternative, comparison), transactional (buy, pricing, download, sign up, near me, a product plus discount), navigational (a brand or site name). When a query could carry two intents, give the primary and secondary and say what would settle it (the SERP format, which I can paste). Mark queries that are a competitor's brand. For each intent, name the page type that usually satisfies it (guide, comparison, product page, category, tool, home page).

Output a table, max 50 rows, columns: keyword, volume, primary intent, secondary intent (or none), confidence (high, medium, low), page type, note. Then a count per primary intent. Do not invent volumes. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the one intent group to build pages for first, no alternatives.

Below is a SERP I copied for one keyword: for each top 10 result the URL, title and page type (article, product, category, tool, video, forum), plus the features present (People Also Ask, video, images, shopping, AI Overview, featured snippet, local pack) in order. If results or features are missing, ask me for them before doing anything.

Count page types and name the dominant format (the type holding 5 or more of the 10; if none, call the SERP mixed and give the top two). Note whether the top 3 share a format the rest do not; whether results are mostly big brands or small sites, which says whether a new page can enter; and which features sit above the first organic result.

Output fixed sections: Shape table (page type, count, positions held), Dominant format, Features above organic, What to build (one page type, one content form: long guide, short answer, list, comparison table, video or tool, and a length band if the pasted pages give it, else "cannot compute from this data"). Never estimate.

Finish with one action: the single page format to build, no alternatives.

Below are: my page's URL and main headings, its target keyword, and the titles and page types of the top 10 results for that keyword. If my page or the SERP is missing, ask me for it before doing anything.

Compare what my page is (type, promise, first 100 words) with what the SERP rewards (the type held by 5 or more results). Mismatch cases: a product page against a SERP of guides; a broad guide against comparisons; a page answering a different question than the query; a locale the SERP does not show. Decide one of: keep, retarget (change the keyword), rebuild (change the page type), split (a second page for the other intent). Do not recommend rebuild when a retarget keeps current traffic; ask for the page's Search Console queries if needed.

Output fixed sections: My page type, SERP type and count, Mismatch (yes or no, evidence quoted), Decision, Reason (max 3 lines). If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single change the decision implies, no alternatives.

Below is a keyword list with volume and my current page for each if any, plus one line on what I sell and who buys it. If either is missing, ask me for it before doing anything.

Map each keyword to a buying stage: problem aware (names a pain, not a solution), solution aware (names a category or approach), product aware (names a product, a comparison, or a brand plus modifier), ready to buy (pricing, trial, buy, discount, near me, a plan name). Then check coverage: which stages have pages, which have volume but no page, and which have a page in the wrong format (a pricing query landing on a blog post). Sum volume per stage from the input only.

Output a table, max 40 rows, columns: keyword, volume, stage, my page (or none), fit (fits, wrong format, missing). Then a stage summary: stage, keyword count, summed volume, pages, gap. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single stage to build for first and the page type, no alternatives.

Below is, per target keyword, the SERP features I saw (featured snippet, People Also Ask, video, image pack, AI Overview, shopping, local pack, top stories), who holds each, and my Search Console position for that keyword. If features or positions are missing, ask me for them before doing anything.

Rules: a featured snippet is winnable when I rank in the top 5 and I can answer more directly than the holder; a People Also Ask question when a page of mine can carry a 40 to 60 word answer under a matching heading; a video slot only if I have a video, so ask; image and shopping packs need a product or original image; a local pack needs a Google Business Profile. Ignore features held by the brand the query names.

Output a table, max 15 rows, columns: keyword, my position, feature, holder, winnable (yes, no, needs asset), what to add (heading, answer block, table, image, video), page. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single feature to target first, no alternatives.

Below is everything on my Google Business Profile: name, categories, description, address, phone, website link, hours, services, attributes, photo count, review count, rating, and last post date. If any field is missing, ask me for it before doing anything.

Audit it. Check: the name has no added keywords (a policy risk); the primary category matches the service earning the most revenue; the description says what, where and for whom in its first 250 characters; every service has its own description; holiday hours are set; there are at least 10 photos and one from the last 90 days; the last post is under 30 days old; with several locations, the website link points to the matching location page, not the home page. If I pasted the category, review count and rating of the top 3 in my map pack, compare against them; if not, write "cannot compute from this data" in those comparisons. Never estimate.

Output a table, max 12 rows, columns: field, current value, problem, fix, priority (high, medium, low).

Finish with one action: the single field to change today, no alternatives.

Below is my canonical name, address and phone (NAP) exactly as it should appear, and a list of listings I found (directory, the name, address and phone shown, and the URL). If the canonical NAP or the listings list is missing, ask me for it before doing anything.

Check consistency field by field. Count a mismatch when: the business name differs beyond case (an added keyword, a dropped Ltd, an old name); the street address differs beyond formatting (Suite versus Ste is formatting; a different number or postcode is a mismatch); the phone number differs beyond formatting; the listing duplicates another listing of the same location. Rank listings: Google Business Profile, Apple Maps, Bing Places, Facebook, industry directories, then the rest.

Output a table, max 25 rows, sorted by importance then by mismatches, columns: directory, URL, name match, address match, phone match, duplicate (yes or no), what to change. Then totals: listings checked, consistent, mismatched, duplicates. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single listing to correct first, no alternatives.

Below are: the service, the area, what makes this location different (team, address, hours, parking, local projects), 2 or 3 real customer quotes from the area, and my existing service page URL. If the local facts or quotes are missing, ask me for them before doing anything; without them this is a doorway page.

Write the location page, under 700 words. Structure: H1 with service and area; a first paragraph under 80 words stating the service, the area covered and one local fact; how the service works here; local proof (the quotes and projects, neighbourhood named); the neighbourhoods served, only the ones I gave; an FAQ of max 4 local questions; a contact block with address, phone and hours. Do not invent projects, quotes, staff names or distances. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output: the page in markdown, then a title tag under 60 characters and a meta description under 155 characters.

Finish with one action: the single local fact I should add to make it stronger, no alternatives.

Below are reviews from my Google Business Profile (stars, text, reviewer first name, date), plus one line on my business and how I sign responses. If either is missing, ask me for it before doing anything.

Write one response per review, max 70 words each. For 4 and 5 stars: thank by first name, echo one specific thing they mentioned, name the service in plain words once, invite them back with something concrete. For 3 stars: thank, acknowledge the specific shortfall, say what changes. For 1 and 2 stars: no defensiveness, acknowledge the specific issue, take it offline with a named person and a channel, never argue facts or disclose customer details. Flag any review that reads as fake or off topic for reporting instead of answering. Never offer discounts in public. Do not invent details about the visit. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output a table, max 20 rows, columns: reviewer, stars, response, flag (respond, report, escalate).

Finish with one action: the single review to answer first, no alternatives.

Below are: my Google Business Profile data (primary category, review count, rating, photo count, website URL, distance from the searched point if known), the same fields for the 3 map pack businesses for my main keyword, and the keyword and location searched. If the competitor rows are missing, ask me for them before doing anything.

Compare field by field: category match to the keyword, review count and rating, review recency if given, photo count, whether the landing page title names the service and city, and whether my address sits inside the searched area. Rank gaps by distance behind (a competitor with 3 times my reviews is a bigger gap than 2 more photos). Mark gaps I cannot close (distance, a category I cannot claim) so I stop chasing them.

Output a table, max 8 rows, columns: factor, me, competitor 1, competitor 2, competitor 3, gap size (large, medium, none), closable (yes or no), what to do. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single closable gap to work on first, no alternatives.

Below are full answers from AI engines (ChatGPT, Perplexity, Gemini, Google AI Overviews, Claude) for a set of prompts, one block per engine and prompt, with the sources each answer showed, plus my domain. If answers, sources or my domain are missing, ask me for them before doing anything.

For each prompt and engine, record whether my domain is cited (a link or a named mention), which page, its position among the sources, and which competitor domains are cited. Count only what is in the pasted answers; do not infer anything about prompts I did not paste. Citation rate per engine is prompts citing me divided by prompts checked.

Output a table, max 30 rows, columns: prompt, engine, my domain cited (yes or no), my page, source position, competitors cited (max 3). Then a summary table: engine, prompts checked, prompts citing me, citation rate, competitor cited most. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single prompt where I am closest to being cited and what to change, no alternatives.

Below are: the prompt, the full text of 2 to 5 competitor pages AI engines cited for it, the passage each engine quoted if I saw it, and my page on the same topic. If any of these is missing, ask me for it before doing anything.

For each cited page find: the passage answering the prompt most directly and how far into its section it sits; whether its heading restates the question; whether it gives a number, definition or step list in one self contained block; date and author signals; structure (table, list or prose). Then compare my page on each point, quoting my closest passage.

Output a table, max 8 rows, one per factor, columns: factor, competitor pages having it (count and which), my page (has it, partial, missing), evidence quoted. Then a Rewrite section: my closest passage rewritten as a 40 to 80 word self contained answer. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Finish with one action: the single change to make to my page, no alternatives.

Below are: the full text of one page of mine, and 3 to 8 questions it should answer (from People Also Ask, question word queries in Search Console, or prompts I tested in AI engines). If the page text or the question list is missing, ask me for it before doing anything.

For each question, write one block of 40 to 80 words that opens with a direct answer sentence that stands alone if quoted, uses the question's own terms, includes one specific fact from my page (a number, a name, a step) and nothing not on the page, and ends without a teaser. Then give its heading (a plain restatement of the question) and its placement (after which existing heading). Mark any question the page cannot honestly answer as "needs new content" instead of writing a block. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output a table, max 8 rows, columns: question, heading, placement, answer block, fact used.

Finish with one action: the single block to add first, no alternatives.

Below are descriptions of my company from different places, each with its source: About page, home page first paragraph, LinkedIn company page, Google Business Profile description, Crunchbase or similar, schema Organization markup, and any Wikipedia or Wikidata entry. If fewer than 3 sources are pasted, ask me for more before doing anything.

Compare across sources: the exact company name and variants, what the company is in one noun phrase (a software company, an agency, a marketplace), founding year, location, founders and titles, product name, category words, and the sameAs links in the schema. Flag every field where sources disagree or one lacks it. Do not fill gaps with facts you do not have. If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output a table, max 10 rows, one per field, columns: field, value per source (abbreviated), consistent (yes or no), canonical value (chosen from the sources, or "ask me"), sources to update. Then the corrected Organization JSON-LD using only confirmed values.

Finish with one action: the single source to correct first, no alternatives.

Below are: my site's URL list with each page's title and meta description, and one line on what the site is and for whom. If titles or the site line are missing, ask me for them before doing anything.

Build an llms.txt file: an H1 with the site name, a one sentence blockquote saying what the site is, then H2 sections each holding a link list (link text, colon, one line description). Include only pages that answer a question or describe a product: skip legal, tag, pagination and author pages, and thin pages (under 200 words when counts are given). Cap at 40 links. Group as Product, Guides, Reference, Pricing, About, plus an Optional section for non core pages. Rewrite descriptions from my meta descriptions, never invent them; where none exists write "needs description". If a number is not in the data I gave you, write "cannot compute from this data". Never estimate.

Output: the llms.txt in a code block, then a table of max 10 skipped pages with the reason.

Finish with one action: the single page whose description I should write first, no alternatives.

# Keyword cluster sheet: [topic]

**Site section:** [URL folder this topic lives in]
**Source of volumes:** [tool name and export date]
**Owner:** [name]

## Clusters

| Cluster | Primary keyword | Supporting keywords | Intent | Target page |
|---|---|---|---|---|
| [cluster name] | [keyword, monthly volume] | [3-8 keywords, comma separated] | [informational / commercial / transactional] | [existing URL or "new"] |
| [cluster name] | [keyword, monthly volume] | [keywords] | [intent] | [URL or "new"] |
| [cluster name] | [keyword, monthly volume] | [keywords] | [intent] | [URL or "new"] |

## Cluster test

One cluster = one page. Two keywords belong together only if the top results for both are mostly the same URLs.

| Keyword pair checked | Shared top-10 URLs | Same page? |
|---|---|---|
| [keyword A] vs [keyword B] | [count out of 10] | [yes / no] |
| [keyword A] vs [keyword C] | [count out of 10] | [yes / no] |

## Pages to create

| New page | Working title | Primary keyword | Priority |
|---|---|---|---|
| [slug] | [title] | [keyword] | [1-3, 1 = first] |

## Fill-in notes

- Volume is a tie-breaker, not a reason to merge two intents onto one page.
- Leave "Shared top-10 URLs" blank if you did not check the results. A blank is more useful than a guess.
- If Claude fills this from an export, it writes "cannot compute from this data" for any intent it cannot infer from the query wording alone.

# Cannibalisation resolution log: [site]

**Export used:** [Search Console performance export, date range]
**Rule for a case:** two URLs each with more than [N] impressions on the same query over [days] days.

## Open cases

| Query | URL A | URL B | Decision | Status |
|---|---|---|---|---|
| [query] | [URL, position 28-day avg, clicks] | [URL, position 28-day avg, clicks] | [keep A / keep B / merge into A / differentiate] | [open / shipped / verified] |
| [query] | [URL, position, clicks] | [URL, position, clicks] | [decision] | [status] |
| [query] | [URL, position, clicks] | [URL, position, clicks] | [decision] | [status] |

## Decision rules

- **Keep one:** the loser gets a 301 to the winner, or its section is folded into the winner and the loser is redirected.
- **Differentiate:** rewrite title, H1 and intro of the weaker page around a different intent, and change internal anchors that point to it with the shared query.
- **Merge:** copy the unique sections of B into A, redirect B, update every internal link to B.

## Fix record

| Query | What shipped | Date shipped | Recheck date | Result after 28 days |
|---|---|---|---|---|
| [query] | [redirect / rewrite / merge, one line] | [date] | [date + 28 days] | [single URL now ranking? position before and after] |

## Fill-in notes

- Position is the 28-day average from the export, never a single-day number.
- A case is closed only when one URL holds the query for 28 days after the ship date.
- If Claude fills this, it uses only rows present in the export and writes "cannot compute from this data" where a page has no impressions on the query.

# Content brief: [primary keyword]

**Target URL:** [URL or planned slug]
**Primary keyword:** [keyword, monthly volume]
**Secondary keywords:** [3-6 keywords that share the same top results]
**Search intent:** [what the searcher wants to have done when they leave the page, one sentence]
**Page type in the top 10:** [guide / listicle / product page / tool / comparison]
**Target length:** [word count of the median ranking page, not the longest]

## Outline

| Section (H2) | Why this section | Source |
|---|---|---|
| [H2] | [cite the ranking page that has it, or delete the section] | [URL of that page] |
| [H2] | [why] | [URL] |
| [H2] | [why] | [URL] |
| [H2] | [why] | [URL] |

## Must include

- **Answer in the first 60 words:** [the direct answer the intro must give]
- **Questions to answer:** [People Also Ask questions seen for the keyword]
- **Proof to show:** [first-hand data, screenshots, examples the writer has access to]
- **Entities to name:** [tools, standards, people the ranking pages all mention]

## Internal links

| Link from this page to | Anchor |
|---|---|
| [URL] | [anchor text] |
| [URL] | [anchor text] |

## Title and meta

- **Title (under 60 characters):** [title]
- **Meta description (under 155 characters):** [description]

## Rules

- Every H2 needs a ranking page in the Source column or it goes.
- Do not add a section because a competitor is longer. Add it because it answers something the searcher asked.

# Page refresh plan: [URL]

**Published:** [date]  **Last updated:** [date]
**Clicks, last 3 months vs same period last year:** [n] vs [n]
**Top query then and now:** [query, position] vs [query, position]

## What changed in the results

| Signal | Then | Now |
|---|---|---|
| Dominant page type in top 10 | [type] | [type] |
| New SERP features | [none / AI overview / video / PAA] | [features] |
| Pages that now outrank us | [URL, why it wins in five words] | [URL, why] |

## Queries to keep, win back, or drop

| Query | Position, 28-day avg | Impressions | Action |
|---|---|---|---|
| [query] | [position] | [n] | [keep / win back / drop] |
| [query] | [position] | [n] | [action] |
| [query] | [position] | [n] | [action] |

## Edits

| Section | Edit | Reason |
|---|---|---|
| [H2 or intro] | [rewrite / add / cut / merge] | [the query or competitor gap it serves] |
| [section] | [edit] | [reason] |

## Facts to update

- [stat or date that is stale, with the current value and its source]
- [screenshot or example to replace]

## Ship and verify

- **Date shipped:** [date]
- **Updated-date changed in the CMS:** [yes / no]
- **Internal links added to this page:** [count, from which URLs]
- **Recheck on:** [date + 28 days]. Compare clicks and position of the "win back" queries.

## Fill-in notes

- "Drop" means stop optimising for it, not delete the text.
- If Claude fills the query table, every position comes from the export. It writes "cannot compute from this data" if the export has no position column.

# E-E-A-T evidence sheet: [URL]

**Topic:** [topic]  **Stakes for the reader:** [low / medium / high]
**Author:** [name]  **Author page:** [URL]

## Experience: proof the author did the thing

| Claim on the page | Evidence shown | Where on the page |
|---|---|---|
| [claim, e.g. "we tested 12 tools"] | [screenshot / data table / photo / dated log] | [section] |
| [claim] | [evidence] | [section] |

## Expertise: why this author

- **Credential or track record relevant to this topic:** [one line, verifiable]
- **Shown on the page as:** [byline / author box / "reviewed by"]
- **Author page links to:** [LinkedIn / publications / professional profile]

## Authority: who else vouches

| External mention | URL | Type |
|---|---|---|
| [site or person that cites this brand on this topic] | [URL] | [citation / interview / listing / award] |
| [mention] | [URL] | [type] |

## Trust: can the reader check us

- **Sources cited in the text:** [count]. Every number links to where it came from: [yes / no]
- **Date and update date visible:** [yes / no]
- **Contact page, about page, editorial policy:** [URLs]
- **Affiliate or sponsor disclosure present where needed:** [yes / no / not applicable]

## Gaps to close before publishing

1. [gap, owner, due date]
2. [gap, owner, due date]

## Rules

- "Evidence shown" must be something a reader can see on the page. A claim without evidence is a gap, not a strength.
- A generic "expert-written" badge is not a credential. Name the person and the reason.

# Title and meta change log: [site]

**Metric source:** [Search Console page export, 28-day window]
**Rule:** one change per page per 28 days, or you cannot tell which edit moved CTR.

## Changes

| URL | Field | Before | After | Date shipped |
|---|---|---|---|---|
| [URL] | [title / meta] | [old text] | [new text] | [date] |
| [URL] | [title / meta] | [old text] | [new text] | [date] |
| [URL] | [title / meta] | [old text] | [new text] | [date] |

## Baseline and recheck

| URL | Clicks, 28 days before | CTR before | Position before | Recheck date |
|---|---|---|---|---|
| [URL] | [n] | [%] | [28-day avg] | [ship date + 28 days] |
| [URL] | [n] | [%] | [avg] | [date] |

| URL | Clicks, 28 days after | CTR after | Position after | Verdict |
|---|---|---|---|---|
| [URL] | [n] | [%] | [avg] | [keep / revert / inconclusive] |
| [URL] | [n] | [%] | [avg] | [verdict] |

## Why each change

- [URL]: [the query the new title targets, and what the old one was missing]
- [URL]: [reason]

## Fill-in notes

- Compare CTR only when position moved less than one place. Otherwise mark "inconclusive" and note the position change.
- Google rewrites titles. On the recheck, note whether the search result showed your text: [yes / no / not checked].
- If Claude fills the recheck rows, it uses the export you paste and writes "cannot compute from this data" for pages missing from it.

# Schema spec sheet: [page template, e.g. product page]

**Applies to URLs matching:** [pattern]
**Schema type:** [Article / Product / FAQPage / LocalBusiness / HowTo / Organization]
**Why this type:** [the visible page content it describes, one line]
**Reference:** [schema.org type URL] and [Google structured data doc URL for this type]

## Properties

| Property | Required by Google? | Source of value | Example |
|---|---|---|---|
| [name] | [yes / recommended] | [CMS field / hard-coded / computed] | [value] |
| [image] | [yes / recommended] | [CMS field] | [URL] |
| [datePublished] | [yes / recommended] | [CMS field] | [ISO date] |
| [author.name] | [yes / recommended] | [CMS field] | [name] |
| [property] | [yes / recommended] | [source] | [example] |

## Rules the developer follows

- Every value must also be visible on the page. If it is not on the page, it is not in the markup.
- Ratings, prices and stock come from the live field, never from a static default.
- One block per page. Nest [linked type, e.g. Organization] inside, do not repeat it.

## Test plan

| Check | Tool | Result | Date |
|---|---|---|---|
| Markup parses with no errors | [Rich Results Test or Schema Markup Validator] | [pass / errors listed] | [date] |
| Values match the rendered page | [manual, 3 sample URLs] | [pass / mismatch listed] | [date] |
| Search Console enhancement report shows the type | [Search Console, after crawl] | [count valid / invalid] | [date] |

## Sign-off

- **Spec approved by:** [name, date]
- **Shipped on:** [date]
- **Recheck enhancement report on:** [date + 14 days]

## Fill-in notes

- "Source of value" is the field that actually exists in the CMS. If the field does not exist, that is a build task, not a placeholder value.

# Technical audit summary: [site]

**Crawl date:** [date]  **Crawler:** [tool]  **URLs crawled:** [n]
**Indexed pages per Search Console:** [n]  **Pages in sitemap:** [n]

## The one constraint

**What is blocking organic growth right now:** [one sentence, e.g. "45% of product pages are not indexed because faceted URLs eat the crawl budget"]

**Evidence:**

- [number from the crawl that shows it]
- [number from Search Console that confirms it]
- [example URL]

**Cost of leaving it:** [pages or clicks affected, and how you estimated it]

## The fix

| Step | Owner | Effort | Done when |
|---|---|---|---|
| [change, one line] | [name or team] | [hours / days] | [observable check] |
| [change] | [owner] | [effort] | [check] |

## Everything else, parked

| Issue | Count | Why it waits |
|---|---|---|
| [issue] | [n] | [smaller than the constraint / needs the fix above first / cosmetic] |
| [issue] | [n] | [reason] |
| [issue] | [n] | [reason] |

## What we will not touch

- [thing the crawler flagged that is fine, and why]

## Recheck

- **Re-crawl on:** [date]
- **Success metric:** [number that must move, and its current value]

## Rules

- If you cannot name one constraint, you have not finished the audit. Do not send a list.
- "Cost of leaving it" uses numbers you have. Write "cannot compute from this data" rather than an estimate with no source.

# Redirect map: [migration or change name]

**Go-live date:** [date]  **Old host:** [domain]  **New host:** [domain]
**Source of old URL list:** [crawl + Search Console pages export + backlinked URLs, date]
**Total old URLs:** [n]  **Mapped:** [n]  **Unmapped:** [n]

## Map

| Old URL | New URL | Type | Rule | Status |
|---|---|---|---|---|
| [old URL] | [new URL] | [301] | [one-to-one / pattern] | [pending / live / verified] |
| [old URL] | [new URL] | [301] | [rule] | [status] |
| [old URL] | [new URL] | [301] | [rule] | [status] |

## Pattern rules

| Pattern (old) | Pattern (new) | URLs covered | Tested on |
|---|---|---|---|
| [/blog/(.*)] | [/articles/$1] | [n] | [3 sample URLs] |

## URLs with no equivalent

| Old URL | Backlinks | Clicks, last 90 days | Decision |
|---|---|---|---|
| [URL] | [n referring domains] | [n] | [redirect to closest parent / 410 / keep live] |

## Must-not-break list

- [URL with the most referring domains]
- [URL with the most clicks]
- [URLs that appear in ads, emails or print]

## Verification after go-live

| Check | Result | Date |
|---|---|---|
| Every old URL returns a single 301 (no chains, no 302) | [pass / list failures] | [date] |
| Old sitemap removed, new sitemap submitted | [done / not done] | [date] |
| Search Console Change of address filed (domain moves only) | [done / not applicable] | [date] |
| Top 50 old URLs by clicks spot-checked in a browser | [pass / failures] | [date] |

## Rules

- Redirect to the page that answers the same intent, not to the home page.
- No redirect points at a URL that itself redirects.

# Index coverage diff: week of [date]

**Property:** [Search Console property]
**Compared:** [this week's date] vs [last week's date]

## Totals

| Status | Last week | This week | Change |
|---|---|---|---|
| Indexed | [n] | [n] | [+/- n] |
| Not indexed, total | [n] | [n] | [+/- n] |
| Sitemap URLs | [n] | [n] | [+/- n] |

## Reasons that moved

| Not indexed reason | Last week | This week | Change | Example URL |
|---|---|---|---|---|
| [Crawled, currently not indexed] | [n] | [n] | [+/- n] | [URL] |
| [Discovered, currently not indexed] | [n] | [n] | [+/- n] | [URL] |
| [Excluded by noindex tag] | [n] | [n] | [+/- n] | [URL] |
| [Page with redirect] | [n] | [n] | [+/- n] | [URL] |

## What explains the change

- [the release, content push or template change that lines up with the date]
- [or: "no deploy this week, movement is Google-side"]

## Pages that should be indexed and are not

| URL | Reason shown | In sitemap? | Internal links to it | Action |
|---|---|---|---|---|
| [URL] | [reason] | [yes / no] | [n] | [add links / fix canonical / request indexing / accept] |

## Action this week

- [one thing, with an owner]

## Fill-in notes

- A change under [5]% of the total is noise. Note it, do not act on it.
- Report numbers lag a few days. Compare the same report on the same weekday.
- If Claude fills this, it subtracts only the numbers you pasted and writes "cannot compute from this data" for any status missing from one of the two weeks.

# Internal link plan: [money page URL]

**Target query:** [primary keyword]  **Position now:** [28-day avg]
**Inlinks today:** [n from crawl, date]  **Click depth from home:** [n]
**Goal:** [n] new contextual links from relevant pages within [n] weeks.

## Current inlinks

| From URL | Anchor text used | In body or in nav/footer? |
|---|---|---|
| [URL] | [anchor] | [body / nav / footer] |
| [URL] | [anchor] | [body / nav / footer] |

## Links to add

| From URL | Why this page | Sentence to place it in | Anchor | Status |
|---|---|---|---|---|
| [URL that mentions the topic and has traffic] | [mentions "[topic]" in section X / ranks for a related query] | [existing sentence, quoted] | [anchor, varies per row] | [to do / done] |
| [URL] | [why] | [sentence] | [anchor] | [status] |
| [URL] | [why] | [sentence] | [anchor] | [status] |
| [URL] | [why] | [sentence] | [anchor] | [status] |

## Anchor mix

| Anchor | Count planned |
|---|---|
| [exact target query] | [at most one third] |
| [partial match or descriptive phrase] | [n] |
| [brand or page name] | [n] |

## Links to remove or fix

- [URL that links with a wrong anchor, or through a redirect]

## Verify

- **Shipped on:** [date]
- **Inlinks after re-crawl:** [n]  **Depth after:** [n]
- **Position at recheck, [date + 28 days]:** [avg]

## Rules

- A link goes in a sentence that already talks about the topic. If you need to write a new sentence just to hold the link, pick another page.
- No two rows use the same anchor for the same target.

# Hub and spoke map: [topic]

**Hub page:** [URL]  **Hub query:** [broad keyword, volume]
**Hub position:** [28-day avg]  **Spokes planned:** [n]  **Spokes live:** [n]

## Spokes

| Spoke URL | Spoke query | Links to hub? | Hub links to it? | Status |
|---|---|---|---|---|
| [URL] | [specific keyword, volume] | [yes, anchor / no] | [yes, anchor / no] | [live / draft / planned] |
| [URL] | [keyword] | [yes / no] | [yes / no] | [status] |
| [URL] | [keyword] | [yes / no] | [yes / no] | [status] |
| [URL] | [keyword] | [yes / no] | [yes / no] | [status] |

## Spoke-to-spoke links

Only where the reader of one would want the other next.

| From spoke | To spoke | Anchor |
|---|---|---|
| [URL] | [URL] | [anchor] |

## Hub page structure

- **Hub intro answers:** [the broad question in 60 words]
- **Each spoke appears in the hub as:** [an H2 summary with a link / a list item / a card]
- **Hub links out to spokes in the body, not only in a sidebar:** [yes / no]

## Missing spokes

| Query with no page | Volume | Nearest existing page | Decision |
|---|---|---|---|
| [keyword] | [n] | [URL] | [new spoke / expand the nearest page] |

## Fill-in notes

- Every spoke links to the hub with a descriptive anchor. The hub links back to every spoke. Missing either direction is an open task, not a preference.
- Spoke count comes from queries that need their own page, not from a target number.
- Recheck hub position on [date + 60 days] and record it here: [avg].

# Link prospect sheet: [campaign or asset]

**Asset we are promoting:** [URL]
**Why someone would link to it:** [one sentence, specific]
**Prospect source:** [competitor backlinks / resource pages / unlinked mentions / broken links, tool and date]

## Prospects

| Site | Page to get the link from | Why they would link | Contact | Stage |
|---|---|---|---|---|
| [domain] | [URL of the page where the link fits] | [they link to a similar resource / broken link on that page / they mention us without a link] | [name, email, role] | [found / contacted / replied / linked / declined] |
| [domain] | [URL] | [reason] | [contact] | [stage] |
| [domain] | [URL] | [reason] | [contact] | [stage] |
| [domain] | [URL] | [reason] | [contact] | [stage] |

## Qualification

| Site | Organic traffic (tool, date) | Topic match | Links out to others? | Keep? |
|---|---|---|---|---|
| [domain] | [n] | [same topic / adjacent / off-topic] | [yes, editorial / no / paid only] | [yes / no] |
| [domain] | [n] | [match] | [yes / no] | [yes / no] |

## Log

| Date | Site | Touch | Result |
|---|---|---|---|
| [date] | [domain] | [email 1 / email 2 / email 3 / social] | [no reply / reply, quoted / linked at URL] |

## Fill-in notes

- "Why they would link" must name a fact about their page. "Great site" is not a reason and gets the row deleted.
- Drop any site where every outbound link looks paid.
- Stop after three touches. Record the result and move on.

# Outreach email, three touches: [prospect site]

**To:** [name, role]  **Their page:** [URL where the link would fit]
**Our asset:** [URL]  **The specific fit:** [what on their page our asset completes or fixes]

## Touch 1, day 0

**Subject:** [their page title or topic, 4-7 words, no "quick question"]

Hi [first name],

On [their page title] you [link to / mention / recommend] [the thing]. [One sentence on what is missing or outdated there, stated as a fact you checked, with the date].

We published [asset title]: [one sentence on what it adds, with a number if you have one].

If it fits, [the exact place on their page where it would go]. If not, no reply needed.

[Your name]
[Role, site]

## Touch 2, day 5

**Subject:** Re: [same subject]

Hi [first name], one more note in case the first got buried. The short version: [their page] [has the gap], [our asset] [fills it]. Happy to send [a summary / the data / an image] if that is easier than reading the page.

[Your name]

## Touch 3, day 12

**Subject:** Re: [same subject]

Hi [first name], last note from me. If [asset] is not a fit for [their page], I will leave it there. If there is someone else who edits that section, I would be grateful for a name.

Thanks either way,
[Your name]

## Rules

- Every bracket is filled with something you checked on their page. If the first paragraph could be sent to any site, rewrite it.
- No attachments, no tracking pixels, no "I hope this finds you well".
- Log every touch and result in the prospect sheet.

# Digital PR pitch: [story working title]

**Headline finding:** [one number and one sentence a journalist can quote verbatim]
**Why now:** [event, season, trend or date that makes it timely]
**Source of the data:** [our data / survey of n people / public dataset, with dates]
**Landing page:** [URL where the full findings live]

## The three facts

1. [finding, number, comparison]
2. [finding, number, comparison]
3. [finding, number, comparison]

## Who covers this

| Outlet | Journalist | Beat | Recent related piece | Angle for them |
|---|---|---|---|---|
| [outlet] | [name] | [beat] | [URL, date] | [which of the three facts, and why it fits their readers] |
| [outlet] | [name] | [beat] | [URL, date] | [angle] |
| [outlet] | [name] | [beat] | [URL, date] | [angle] |

## Assets ready

- [chart, image or table, file link]
- [expert quote, name and role, approved]
- [methodology note, URL]

## Pitch email

**Subject:** [the headline finding in under 60 characters]

Hi [first name],

[Headline finding in one sentence]. [Second sentence: the comparison or the surprise].

Full data and method: [URL]. [Expert name] can comment today.

[Your name, role, phone]

## Fill-in notes

- If the headline finding needs a paragraph to explain, it is not the headline finding.
- Every number links to a page a journalist can check. No page, no pitch.
- Log sends and replies: [date, outlet, result].

# Monthly SEO report: [month, year]

**Property:** [site]  **Prepared by:** [name]  **Date:** [date]

## Headline

[One sentence: what happened and the single reason, e.g. "Clicks up 12%, driven by three refreshed pages"]

## Totals

| Metric | This month | Last month | Same month last year |
|---|---|---|---|
| Clicks | [n] | [n] | [n] |
| Impressions | [n] | [n] | [n] |
| Average CTR | [%] | [%] | [%] |
| Average position | [avg] | [avg] | [avg] |

## What moved

| Page | Clicks change | Why |
|---|---|---|
| [URL] | [+/- n] | [refresh shipped on date / new page / lost a SERP feature / seasonal] |
| [URL] | [+/- n] | [reason] |
| [URL] | [+/- n] | [reason] |

## What we shipped

| Item | Date | Expected effect | Seen yet? |
|---|---|---|---|
| [page, fix or link batch] | [date] | [metric it should move] | [yes / no / too early] |

## Next month

1. [one action, owner, due date]
2. [one action, owner, due date]

## Risks

- [thing that could drop traffic next month, and what we are doing about it]

## Fill-in notes

- Year-on-year is the comparison that matters for seasonal sites. Month-on-month explains what we did.
- "Why" must be a thing we did or a thing we observed in the results, not "algorithm update" without a page-level example.
- If Claude fills the tables, every number comes from the pasted exports and it writes "cannot compute from this data" for any cell the exports do not cover.

# Quick-win tracker: [site]

**Export:** [Search Console pages export, 28 days ending date]
**Position 8-15 rule:** impressions above [n], position between 8 and 15.
**Low-CTR rule:** position under [5], CTR below [%].

## Candidates

| URL | Query | Position, 28-day avg | Impressions | CTR |
|---|---|---|---|---|
| [URL] | [top query for the page] | [avg] | [n] | [%] |
| [URL] | [query] | [avg] | [n] | [%] |
| [URL] | [query] | [avg] | [n] | [%] |
| [URL] | [query] | [avg] | [n] | [%] |

## Action per page

| URL | Action | Date shipped | Recheck date |
|---|---|---|---|
| [URL] | [title rewrite / add answer block / 3 internal links / refresh section X] | [date] | [date + 28 days] |
| [URL] | [action] | [date] | [date] |
| [URL] | [action] | [date] | [date] |

## Results

| URL | Position after | Clicks before | Clicks after | Verdict |
|---|---|---|---|---|
| [URL] | [avg] | [n] | [n] | [win / no change / worse] |
| [URL] | [avg] | [n] | [n] | [verdict] |

## Fill-in notes

- One action per page per cycle. If you change the title and add links the same week, you will not know which one worked.
- Search Console's interface filters by dimension (page, query, country, device). The position and CTR cut-offs are applied in the export, in a spreadsheet.
- If Claude fills the candidates, it applies only the two rules above to the pasted rows and writes "cannot compute from this data" where a column is missing.

# Traffic drop incident note: [site], [date noticed]

**Drop first visible on:** [date, from the Search Console date chart]
**Size:** [clicks per day before] to [clicks per day after], [%] change
**Scope:** [whole site / one folder / one page type / one country / one device]

## What we checked

| Check | Result |
|---|---|
| Deploys or CMS changes in the 14 days before the drop | [none / list with dates] |
| Robots, noindex, canonical changes | [none / what changed] |
| Pages report: not-indexed counts moved? | [no / which reason, by how much] |
| Manual actions or security issues | [none / details] |
| Google update announced around the date | [none / name and date] |
| Same period last year (seasonality) | [similar dip / no dip] |

## Pages that lost the most

| URL | Clicks before | Clicks after | Top query lost | Position before and after |
|---|---|---|---|---|
| [URL] | [n] | [n] | [query] | [avg] to [avg] |
| [URL] | [n] | [n] | [query] | [avg] to [avg] |
| [URL] | [n] | [n] | [query] | [avg] to [avg] |

## Current hypothesis

[One sentence. What most likely caused it, and the evidence line above that supports it.]

## Decision

- [ ] Wait and re-measure on [date + 14 days], no changes
- [ ] Revert [specific change] on [date]
- [ ] Fix [specific issue], owner [name], by [date]

## Follow-up

- **[date + 14 days]:** [clicks per day now, hypothesis holds / rejected]
- **[date + 28 days]:** [status, closed or escalated]

## Rules

- Fill the checks table before writing a hypothesis.
- Do not ship a fix and a revert in the same week.

# SERP analysis: [keyword]

**Searched on:** [date]  **Country:** [country]  **Device:** [mobile / desktop]  **Logged out, incognito:** [yes / no]
**Monthly volume:** [n, source]

## Shape

| Feature present | Position on the page | What it contains |
|---|---|---|
| [AI overview] | [top / none] | [which sites it cites] |
| [People also ask] | [after result n] | [the questions, verbatim] |
| [video / images / shopping / local pack / featured snippet] | [position] | [what shows] |

**Dominant intent:** [informational / commercial / transactional / navigational]
**Dominant page type:** [n of 10 are guides / listicles / product pages / tools / forums]

## Who ranks and why

| Rank | URL | Page type | Why it ranks |
|---|---|---|---|
| 1 | [URL] | [type] | [the thing it has that others do not: data, tool, brand, depth, freshness] |
| 2 | [URL] | [type] | [why] |
| 3 | [URL] | [type] | [why] |
| [4-10 summarised] | [pattern] | [pattern] | [pattern] |

## Common elements across the top 5

- [section or element present in most of them]
- [format element: table, calculator, video, comparison]
- [oldest and newest publish dates seen]

## What a page must do to enter

- **Match the page type:** [type]
- **Answer in the first screen:** [the direct answer or the thing the searcher wants to see]
- **Have what the winners have:** [element]
- **Add what none of them have:** [gap, only if the searcher would want it]

## Fill-in notes

- Fill "why it ranks" from reading the page, not from a tool score.
- If the top 10 is split between two page types, this is two keywords. Pick one intent and note the other here: [other intent].

# Buying-stage content map: [product or service]

**Audience:** [who buys this, one line]
**Keyword source:** [tool, export date]  **Page source:** [crawl or CMS export, date]

## Map

| Stage | Question the searcher has | Query examples | Page that answers it | Gap? |
|---|---|---|---|---|
| Problem aware | [what is going wrong for them] | [2-3 queries, volume] | [URL or none] | [yes / no] |
| Solution aware | [what kinds of fixes exist] | [queries] | [URL or none] | [yes / no] |
| Comparing | [which option, which vendor] | [queries: "vs", "best", "alternatives"] | [URL or none] | [yes / no] |
| Ready to buy | [price, how to start, trust] | [queries: "pricing", "demo", "reviews"] | [URL or none] | [yes / no] |
| Using | [how to do X with the product] | [queries] | [URL or none] | [yes / no] |

## Stage handoff

| From page | Links to next-stage page | Anchor |
|---|---|---|
| [problem-aware URL] | [solution-aware URL] | [anchor] |
| [comparing URL] | [ready-to-buy URL] | [anchor] |

## Gaps ranked

| Stage | Missing page | Volume of the queries it would serve | Priority |
|---|---|---|---|
| [stage] | [working title] | [n] | [1-3] |

## Fill-in notes

- One query per stage is enough to prove the stage exists. Volume decides priority between gaps, not which stage gets a page.
- A page only counts as answering a stage if its top query in Search Console belongs to that stage. Otherwise it is a gap with a page in the way.
- Recheck this map every [quarter] and note the date: [date].

# Google Business Profile audit: [business name, location]

**Profile URL:** [Maps URL]  **Audited by:** [name]  **Date:** [date]
**Primary category:** [category]  **Verified:** [yes / no]

## Core fields

| Field | Current value | Matches website? | Action |
|---|---|---|---|
| Business name | [as shown] | [yes / no] | [none / fix: real name only, no keywords] |
| Address | [as shown] | [yes / no] | [action] |
| Phone | [as shown] | [yes / no] | [action] |
| Website URL | [as shown] | [links to location page? yes / no] | [action] |
| Hours, including holidays | [as shown] | [yes / no] | [action] |
| Primary category | [category] | [same as top competitor? yes / no] | [action] |

## Content

| Item | Status | Action |
|---|---|---|
| Description states what and where | [yes / partial / no] | [action] |
| Services or products listed with descriptions | [n listed] | [action] |
| Photos: exterior, interior, team, work, logo | [count per type] | [add which] |
| Posts in the last 30 days | [n] | [action] |
| Q&A answered by the owner | [n open / n answered] | [action] |

## Reviews

- **Count and average:** [n, rating]
- **Received and replied to, last 90 days:** [n received, n replied]
- **Top competitor count and rating:** [n, rating]

## Consistency check

| Directory | Name | Address | Phone | Match? |
|---|---|---|---|---|
| [website footer] | [value] | [value] | [value] | [yes / no] |
| [Apple Maps / Bing Places / main industry directory] | [value] | [value] | [value] | [yes / no] |

## Top 3 fixes

1. [fix, owner, date]
2. [fix, owner, date]
3. [fix, owner, date]

## Rules

- Copy values as shown on the live profile, not what they should be.

# Location page spec: [service] in [city or area]

**URL:** [/locations/city-slug]  **Target query:** [service + city, volume]
**Linked Google Business Profile:** [Maps URL or "none, service area business"]

## Facts only this page has

| Fact | Value | Source |
|---|---|---|
| Address and phone for this location | [value] | [profile / office manager] |
| Opening hours | [value] | [source] |
| Team members at this location | [names, roles] | [source] |
| Areas served from here | [neighbourhoods, towns] | [source] |
| Local proof: reviews, jobs done, clients | [3 examples with dates] | [source] |
| Local specifics of the service | [regulation, climate, pricing difference, common local issue] | [source] |

## Page sections

| Section | Content | Unique to this location? |
|---|---|---|
| H1 | [service in city] | [yes] |
| Intro, 60 words | [what we do here and for whom] | [yes] |
| Local proof | [reviews or case examples from the table above] | [yes] |
| Service details | [what is included, priced for this area] | [partly] |
| Map and directions | [embedded map, landmarks, parking] | [yes] |
| FAQ | [3-5 questions customers here actually ask] | [yes] |
| Call to action | [phone, form, booking link] | [shared] |

## Technical

- **Title:** [service in city, brand, under 60 characters]
- **Schema:** [LocalBusiness with this address, phone, hours, geo]
- **Internal links from:** [service page, locations index, nearest other location page]
- **Photos:** [n, taken at this location, alt text naming the place]

## Fill-in notes

- Fewer than four rows filled in the facts table means the page is not ready. Collect the facts first.
- Never reuse a paragraph from another location page with the city name changed.

# AI citation tracker: [brand], [month]

**Questions tested:** [n]  **Engines:** [list, e.g. AI overviews, ChatGPT, Perplexity, Copilot]
**Run on:** [date]  **Logged out and no history:** [yes / no]  **Country:** [country]

## Results

| Question | Engine | We are cited? | Who is cited | Our page that should be |
|---|---|---|---|---|
| [question as typed] | [engine] | [yes, URL / no] | [domains listed in the answer] | [URL] |
| [question] | [engine] | [yes / no] | [domains] | [URL] |
| [question] | [engine] | [yes / no] | [domains] | [URL] |
| [question] | [engine] | [yes / no] | [domains] | [URL] |
| [question] | [engine] | [yes / no] | [domains] | [URL] |

## Summary

| Engine | Questions run | Cited in | Rate |
|---|---|---|---|
| [engine] | [n] | [n] | [%] |
| [engine] | [n] | [n] | [%] |

**Change vs last month:** [+/- n citations, which questions flipped]

## Who wins the questions we lose

| Domain | Times cited | What their cited page has | Our nearest page |
|---|---|---|---|
| [domain] | [n] | [a table, a definition in the first paragraph, a dated stat, an original study] | [URL] |
| [domain] | [n] | [what] | [URL] |

## Action

- [one page to change, what to add, based on the column above]

## Fill-in notes

- Type each question exactly the same way every month. Keep the list in this file.
- Record what the answer shows on screen, not what you expected.
- A single run is a sample. Note in the summary if you ran each question more than once: [runs per question].

# Entity fact sheet: [brand]

**Owner:** [name]  **Last checked:** [date]  **Canonical source:** [about page URL]

## Core facts

| Fact | Value | Where it must appear |
|---|---|---|
| Legal name | [value] | [about page, footer, Organization schema, LinkedIn] |
| Brand name as written | [value, exact casing] | [everywhere] |
| One-line description | [what it is, for whom, in 20 words] | [about page, schema description, social bios, directories] |
| Category | [the category term engines use for this kind of company] | [about page, schema, Wikipedia-style sources] |
| Founded | [year] | [about page, schema] |
| Founders and key people | [names, roles] | [about page, author pages, LinkedIn] |
| Headquarters | [city, country] | [about page, schema, profiles] |
| Logo | [URL of the canonical file] | [schema, profiles] |

## Where the facts live

| Location | Matches this sheet? | Last checked | Fix |
|---|---|---|---|
| [about page] | [yes / no, which fact differs] | [date] | [action] |
| [Organization schema on the home page] | [yes / no] | [date] | [action] |
| [LinkedIn company page] | [yes / no] | [date] | [action] |
| [Crunchbase / Wikidata / industry directory] | [yes / no] | [date] | [action] |

## Same-as links

- [LinkedIn URL]
- [Wikidata or Crunchbase URL]
- [other profile URLs]

## Facts engines currently get wrong

| Wrong fact seen | Engine | Likely source of the error | Fix |
|---|---|---|---|
| [what the answer said] | [engine] | [stale directory / old press / no source at all] | [update the source / publish a page stating the fact] |

## Rules

- One value per fact. If two teams use different descriptions, decide here and update both.
- Change this sheet first, then the locations. Never the other way round.

# [Brand name]

> [One or two sentences: what the site is, who it is for, what it is known for. Plain facts, no slogans.]

[One short paragraph of context an engine needs to interpret the pages below: the product or service, the main topics covered, the languages available, the country focus.]

## [Core pages]

- [Page title](https://[domain]/[path]): [one line on what the page states or answers]
- [Page title](https://[domain]/[path]): [one line]
- [Page title](https://[domain]/[path]): [one line]

## [Docs or guides]

- [Guide title](https://[domain]/[path]): [what question it answers]
- [Guide title](https://[domain]/[path]): [what question it answers]
- [Guide title](https://[domain]/[path]): [what question it answers]

## [Data and research]

- [Study title](https://[domain]/[path]): [the headline finding with its number and date]

## [About]

- [About](https://[domain]/about): [legal name, founded year, location, founders]
- [Contact](https://[domain]/contact): [how to reach the company]

## Optional

- [Changelog or blog index](https://[domain]/[path]): [updated how often]
- [Pricing](https://[domain]/pricing): [plans listed, currency]

## Fill-in notes

- Keep the file to the pages you would hand a new employee on day one. Fewer than [40] links. If a page is not worth reading, it is not worth listing.
- Each description is a fact about the page, written so it stands alone.
- Match every fact here against the entity fact sheet before publishing.
- The "Optional" heading is the only section an engine may skip. Delete this notes block before publishing.