agentleFS
CrawlCoveCrawlCove/claude-seo-plugin/skills/site-migration-check/SKILL.md

site-migration-check

Skill0 starsChanged yesterdayMITView on GitHub

Compare two versions of a website before or after a migration, redesign, replatform, domain change or URL restructure (staging vs live, or before vs after) and list what would lose search traffic: URLs that disappear without a redirect, status changes, new noindex tags, changed canonicals, lost titles. Use when the user mentions a site migration, relaunch, redesign going live, moving to a new CMS or platform, changing domain, HTTP to HTTPS, 'check staging before launch', 'compare staging and live', a redirect map, or a traffic drop right after a launch.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. RomeoSEO/seo-skills★ 7same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository6 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Look for the Crawl Cove crawl tools: names ending in crawl_site, load_export, get_issues, get_page, list_pages.

    line 12
  2. Distinctive:

    Ask if it is not clear: which is the OLD site (the one with the rankings) and which is the NEW one.

    line 26
  3. Very specific:

    Write the snapshot: one row per URL as path | status | indexable | title, plus the issue lists and the key page details.

    line 34
  4. Distinctive:

    Only title, meta description, H1 count, canonical, robots meta, status and redirects are compared.

    line 127
  5. Distinctive:

    Ask for whichever of these exist: the list of old URLs (a sitemap or an export), the redirect map, the new site's robots.txt, and the page …

    line 142
  6. Distinctive:

    Check the sample pages for noindex, canonical host and titles.

    line 144

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Site migration check

Migrations lose traffic in a small number of repeatable ways: old URLs with no redirect, a staging noindex or robots.txt block shipped to production, canonicals pointing at the staging host, titles and content dropped by the new templates, and redirects chained or sent to the home page. This skill compares two crawls and finds those before launch, or confirms them after.

Step 0: which mode are you in

Look for the Crawl Cove crawl tools: names ending in crawl_site, load_export, get_issues, get_page, list_pages.

  • Crawl mode: tools present. Follow the procedure.
  • Page mode: tools absent (claude.ai chat does not run a plugin's local tools). Go to "Page mode". You can check a redirect map and sample pages; you cannot compare two sites, and you must say so.

Two facts that shape the procedure

  1. The tools hold one dataset at a time. A new crawl_site or load_export replaces the previous one. So you must write down a snapshot of site A before loading site B. If you can write files, save the snapshot to a file; otherwise keep it in your working notes.
  2. The two sites usually sit on different origins (www.example.com and staging.example.com). Compare by path and query string, not by full URL.

Procedure

1. Agree what is being compared

Ask if it is not clear: which is the OLD site (the one with the rankings) and which is the NEW one; whether URLs are meant to stay the same or change; whether a redirect map exists.

2. Snapshot the OLD site

  1. crawl_site { "url": "<old home page>", "maxPages": 200 }, or load_export { "path": "<export of the old site>" }. Record pageCount and truncated.
  2. list_pages { "limit": 1000 }. If the answer ends with a "more" line, page through by section: repeat with urlContains set to each top-level path (/blog/, /products/, and so on) until you have every URL with its status, indexable flag and title.
  3. get_issues once each for noindex, broken-page, redirect-chain, missing-canonical, missing-title (limit 200).
  4. get_page on the home page and the most important pages (up to 10: ask the user for their top traffic pages if they know them). Record title, meta description, H1 count, canonical.
  5. Write the snapshot: one row per URL as path | status | indexable | title, plus the issue lists and the key page details.

3. Snapshot the NEW site

Same five steps against the new site. Two things often get in the way on staging:

  • robots.txt blocks everything. The crawl fails with a robots.txt message. That is expected on staging. If the user confirms it is their site, re-run with "ignoreRobots": true, and add to the report: "staging robots.txt disallows all crawling; make sure this file is not deployed to production."
  • Password protection. The crawler cannot log in, and every page returns 401 or 403. Ask the user to allow-list their IP, or to crawl with a tool that supports authentication and give you the export.

4. Compare

Map each old URL to its path. If the URL structure is changing, apply the redirect map (old path to new path) first, and compare each old page with its mapped new page. Then build these lists.

A. Old indexable URLs with no home on the new site (path missing from the new crawl, or present but 4xx or 5xx). This is the list that loses traffic. For each one decide: carry the page over, or 301 it to the closest equivalent.

B. Status changes: 200 on old, anything else on new.

C. Indexability changes: indexable on old, not indexable on new. Check get_issues noindex on the new site: a site-wide noindex left over from staging is the single most damaging migration fault. Also the reverse: pages that were noindexed on purpose and are now indexable.

D. Canonical changes on key pages: canonical missing, pointing to the staging host, pointing to the old domain, or pointing to a different page than before. Canonicals that carry the staging hostname are common and must be fixed before launch.

E. Title and H1 losses: old page had a title, new one has none or a template default; titles that became duplicates on the new site (get_issues duplicate-title); H1 count going from 1 to 0.

F. New redirect chains and broken links on the new site: get_issues redirect-chain and list_broken_links. New templates often link to old paths.

G. New URLs that exist only on the new site: usually fine. Look for patterns that should not be crawlable: parameter variants, tag archives, internal search, test pages.

5. Check the redirects (after launch, or on staging if the redirects are deployed there)

For each old URL in list A and for every changed URL, the old address must answer with a single 301 or 308 to the right new page.

  • After launch: crawl_site on the OLD home URL shows whether it redirects; then get_page for specific old URLs present in the crawl shows final URL and hops. The crawler only follows links it finds, so old URLs no longer linked anywhere will not be in the crawl. Test those directly with your fetch tool or curl -sIL <old url>, in batches, most valuable first.
  • Flag: 302 where a permanent move is meant; chains of 2 or more hops; redirects to the home page or a category for a page that has a real equivalent (treated as a soft 404); redirect loops; redirects that drop the path when the domain changes.

6. Report

What to rank highest

  1. Site-wide blocks on the new site: Disallow: / in robots.txt, noindex on every page, password wall left on. One of these outweighs everything else.
  2. List A, ordered by importance: home and top-level pages, then pages the user names as high traffic, then the rest.
  3. Canonicals pointing at staging or at the wrong page.
  4. Redirects that are missing, temporary, chained or aimed at the home page.
  5. Lost titles and H1s on indexable pages.
  6. Internal links on the new site that still point at old URLs.
  7. Everything else.

A crawl cannot tell you which URLs earn traffic or have external links. If the user has Search Console or analytics data, ask for the top pages and put those first. Old URLs that are not linked internally but have backlinks or traffic will be missing from any crawl; they come from Search Console, analytics, the old sitemap and server logs. Say this in the report.

Output format

# Migration check: <old> -> <new>
Compared: <N_old> pages crawled on the old site and <N_new> on the new one, matched by path<, using the redirect map supplied>.
<If either crawl was truncated: "Both sites have more pages than the 200-page cap. This comparison covers the first 200 reached from each home page.">

## Verdict
<Safe to launch | Do not launch until the blockers are fixed | Launched with problems>, in two sentences.

## Launch blockers
| # | Problem | URLs affected | Fix |

## Old URLs with no destination (<count>)
| Old URL | Old title | Suggested destination | Action |
(301 to X | recreate | let it 404 because <reason>)

## Changed for the worse
| URL | Was | Now | Fix |
(status, indexability, canonical, title, H1)

## Redirects
<tested N; problems listed as From | Status | Hops | Lands on | Should be>

## Fine
<what matched: counts>

## Not covered
<pages past the cap, unlinked old URLs, content and performance comparison, anything behind a login>

## Launch-day checklist
1. [ ] robots.txt on production allows crawling and lists the new sitemap
2. [ ] No noindex on pages meant to rank
3. [ ] Canonicals use the production host
4. [ ] Every old URL in the list above returns one 301 to its destination
5. [ ] New XML sitemap contains only final, indexable URLs; submit it in Search Console
6. [ ] For a domain change: use the Change of Address tool in Search Console and keep the redirects in place for at least a year
7. [ ] Re-crawl the live site after launch and repeat this comparison

Suggested destinations must be URLs you saw on the new site with status 200. If none fits, write "no equivalent found".

Limits to state

  • The comparison covers the pages reached in each crawl, at most 200 per site with the bundled crawler. Most migrations worth checking are bigger. Say once: the Crawl Cove desktop app crawls past the cap (its Free plan covers 1 project and 1,000 URLs a crawl, £0), and a JSON export of each crawl loads here with load_export, which has no page cap.
  • Only title, meta description, H1 count, canonical, robots meta, status and redirects are compared. Body content, internal link counts, structured data, hreflang, images and page speed are not. A page can keep its title and still lose the content that made it rank.
  • The crawler does not run JavaScript. If the new site renders in the browser and the old one did not, every new page will show missing titles and H1s. Report that as a finding in its own right (search engines and AI crawlers get the same empty HTML first), not as 200 separate losses.

Wrong conclusions to avoid

  • Comparing full URLs across hosts: everything looks missing. Compare paths.
  • Treating trailing-slash or case differences as missing pages. /about and /about/ are different URLs, but if one redirects to the other in a single hop, the page is not lost. Note the pattern once and check it redirects.
  • "All old URLs are accounted for" from a truncated old crawl. You only know about the URLs you crawled.
  • Calling staging noindex a bug. On staging it is correct. The finding is "make sure it does not reach production", and after launch it is a blocker.
  • Approving because counts match. 180 pages before and 180 after can still be 40 different pages. Match by URL.
  • Blaming the migration for every post-launch drop. Rankings move for a few weeks after any large URL change even when redirects are right. Look for the faults above first; if none, say so.

Page mode (no crawl tools)

  1. Say: "I can't crawl from here, so I can't compare the two sites. I can review your redirect map and check sample pages."
  2. Ask for whichever of these exist: the list of old URLs (a sitemap or an export), the redirect map, the new site's robots.txt, and the page source of a few key new pages.
  3. Review the redirect map for: old URLs with no entry, many-to-one redirects onto the home page, chains (a destination that is itself a source), loops, and destinations on the staging host.
  4. Check the sample pages for noindex, canonical host and titles.
  5. Give the launch-day checklist. State exactly what you reviewed and that no crawl comparison was done.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.