Website launch check
A conversation, not a static report: settings first, then the site's own tests, then the
checks the tests cannot do, a question only where the owner's intent decides, an offer to fix
each gap, and a dated report. The skill's text is English; everything the owner sees is in the
owner's language (§1).
When it runs:
| Run |
When |
Report |
| baseline |
new-website step 5a: after the scaffold is green (and after the brand-kit intake, where there is one) |
launch-check-<date>-baseline.md, verdict "Baseline, not a launch verdict" |
| pre-launch |
new-website step 10c: after steps 7 to 10b, before §4's deploy |
launch-check-<date>-prelaunch.md, verdict 🟢 / 🟡 / 🔴; §4 starts after 🟢, or a 🟡 the owner accepts |
| live |
new-website step 12a: once launch day's work is done (the production build Active on the live domain, the domain attached); not optional, since the zone read (§3) runs only here for a new domain |
launch-check-<date>-live.md |
It also runs on its own, on any site built with this toolkit, whenever the owner asks.
0. Settings: docs/launch/config.json
Read docs/launch/config.json. If the site has an older .launch-check-config.json in its
root, move it there first: git mv when git tracks it, mv when not, and say so. If both
exist, move nothing and ask which to keep. Make the folder first (mkdir -p docs/launch). Older reports in docs/ stay where they are. The
old file's keys storybrand and contact_forms become story_layer and contact_form; fill a
missing key as below. A moved file counts as saved settings (the last paragraph of this section).
Without a settings file, fill one from what the site already says, and ask only what is
still open:
| Setting |
From |
business_in_germany (where the business is based decides the legal checks, not the site's language) |
ask; the interview does not ask it |
site_languages |
SITE.locale in src/config.ts, the README's interview answers (Q4) |
story_layer |
the README's record of the §2a story offer, STORY.md; never ask again |
blog, contact_form |
the README's interview answers (Q1, Q2), the pages |
testimonials |
src/data/testimonials.ts (from website-testimonials); ask if the site has none |
analytics |
USE_ANALYTICS and the README (Q5) |
report_language |
the owner's language (§1) |
replaces |
ask once: does this site replace one that is live now? Its address, or null |
{
"business_in_germany": true,
"site_languages": ["de"],
"story_layer": false,
"blog": false,
"testimonials": true,
"contact_form": true,
"analytics": "plausible",
"report_language": "de",
"replaces": "https://www.example.com"
}
On later runs, name the saved settings in one line and ask once whether any changed.
1. Language
Questions, findings, reports, todos and the HTML are in the language the owner writes in;
report_language is the default on a re-run, and the owner's message overrides it. A German
owner gets everything in German, headings included. This file stays English.
2. The site's tests first
Run, in the site's folder, npm run build, then the whole suite with
npx playwright test --forbid-only, keeping the whole output. Use it, not npm test, so a
narrowed test script cannot hide a failure, and --forbid-only so a test.only left in a spec
fails the run instead of narrowing it; if package.json's test script is anything but playwright test, that is
a 🟡 finding. Then follow references/test-coverage.md:
- A failing test is a 🔴 blocker that names the test and quotes its message. No look by eye
stands in for a test.
- Every
⚠ placeholders: line is a blocker (a page still listed in UNFILLED_UNTIL_LAUNCH,
or a [MISSING: …] placeholder, which only warns outside CI). ⚠ positioning coverage: is a
🟡 finding. A skipped test is a 🟡 finding.
- A category the mapping gives to a test is left to that test, per page, only where the
output shows that page's test passed. The report lists those under "Left to the site's tests,
as they stand in this site, for the pages in
PAGES", and the check does not look at them
again on those pages. Every other page, and every built page that PAGES lacks, is the
check's.
3. What the launch check checks
Only what the tests cannot. Skip what the settings rule out, and list it under "Skipped by the
settings".
Always:
- Titles and descriptions: no two pages share one, and each says what its page is (🟡).
- What no test reads (🔴, since a visitor would read it): the starter's example values (the name
Example, hello@example.com) on the home page and the legal pages (the imprint where the
site has one, the privacy or Datenschutz page), public/llms.txt's own
[bracket] slots, the 404 page's text.
- Console errors on the local preview (
npm run preview), read with the site's own Playwright (a
short script that opens each built page and collects console errors). Run it after the tests,
and stop it when done: the tests start a preview of their own, and Astro refuses it while
another one runs. Errors that come from
the _headers rules, HTTPS and the 404 status show only on a deployed site: at the
baseline and pre-launch runs they read "not checked yet: not deployed", and the live run
checks them.
- The legal pages' answer (
README.md in the folder that holds package.json): quote, in the
report's Imprint / privacy row, every line in one of the three forms AGENTS.md §5 gives
(legal pages: reviewed by a lawyer (<date>), legal pages: own texts (<date>), legal pages: published as drafted (owner's decision, <date>)), and ask the owner about them at every
pre-launch run, as below.
- No such line: a blocker (🔴): the owner's answer is not recorded;
AGENTS.md §5 asks for it.
Do not ask here. Claim no test failure: the site's tests pass a legal page still listed in
UNFILLED_UNTIL_LAUNCH whose draft notice shows.
- Which pages each line covers: the README names none. With one line, it covers every legal
page. With more than one, ask the owner at each pre-launch run which legal pages each line
was recorded for. A page's answer is the newest line that covers it (of two with one date,
the one further down the README); quote older lines as history, with no question.
- A page whose answer is
reviewed by a lawyer or own texts: at each pre-launch run, quote
that line and ask the owner whether it is still true for the pages it is the answer for, one
question per line. "Yes" goes in the report as "owner confirmed, ". "No" or "don't
know" is a blocker that points at AGENTS.md §5: record a true answer, then launch.
- A page whose answer is
published as drafted: no question. Report it as 🟡, one per such
line, naming the pages it is the answer for and leaving out any the date check below
blocks, and none for a line left with no pages: ": the kit's drafts, published on
your decision of ; a lawyer's review is still open to you."
- A legal page no line covers, or added after its line: at every pre-launch run, for each
legal page in
tests/_helpers.ts PAGES (last path segment privacy, datenschutz,
legal-notice or impressum), run
git log --diff-filter=A --format=%cs -- <its file under src/pages/>, one file per call,
and take the first line: the date it last entered git under that name. Not --follow,
which dates a renamed file (the /datenschutz or /impressum swap) by the file it came
from. A page rewritten under the same file name keeps its first date, so this check cannot
see the rewrite. A page that no line covers, or whose date is after its answer's date, is a blocker
naming that page: AGENTS.md §5 asks again for it and adds a dated line. The answer's own
day counts as covered. No date (no output: the file is not committed; or an error: not a
git repo) is a blocker of its own: commit the page, then run the check again. This holds
whatever the answer, published as drafted included. Claim no test failure: the tests
compare no dates.
Business in Germany:
- Legal Notice or Impressum per § 5 DDG (the Digitale-Dienste-Gesetz replaced the TMG in May 2024).
The kit ships the English page at
/legal-notice; the German one at /impressum replaces it on a
German-language site, and an owner based in Germany may route both. Either address counts as the
imprint. It carries: name,
address, a fast electronic contact, register entry, VAT ID if any; for a regulated
profession (architects, lawyers, doctors, and a trade that must register with a chamber such
as the Handwerkskammer) also the chamber, the professional title and the state that granted it.
- Datenschutz: hosting, analytics, forms, any embeds, the contact for data questions.
Whether the text fits what the site really does; the slots and the analytics settings are
new-website §4's own item.
- Consent banner: needed only if the site stores or reads non-essential data on the
visitor's device (§ 25 TDDDG, formerly TTDSG), such as Google Analytics, YouTube or Google
Maps embeds. Analytics that store and read nothing on the device need no § 25 consent (the
privacy page still names their legal basis, which is the owner's to settle), but check it on
the deployed site (the browser's storage panel after a page view) rather than
assume it from the product's name. If anything is stored or read, or you cannot tell, report
the banner question as open for the owner, not as passed. A banner the site does not need
is itself a finding.
Business elsewhere: a privacy policy (GDPR-ready where the site serves people in the EU),
and terms where the site sells or takes bookings.
Story layer: the home page against STORY.md (from website-story), not against a
generic list.
Blog: an index page, at least one post (or drafts ready), a feed, a named author.
Testimonials (src/data/testimonials.ts): real names and roles where the owner chose names, and consent on file for
each named person: at each pre-launch run, ask the owner, quote by quote. "Yes" goes in the report as "owner confirmed
consent on file, ". "No" or "don't know" for a quote is a blocker that points at
website-testimonials' consent rule (keep it off the live site until consent lands). Keep no
consent record and no contact details.
Contact form: it shows its confirmation; a privacy note beside it links the privacy page;
spam protection where it sends to a server. Ask before sending a test message: on a live site
it reaches a real inbox.
E-E-A-T and local signals (an assessment, not a measurement): rate Experience, Expertise,
Authority, Trust and Local from what the site shows, 1 to 5 with one line why. Authority
needs outside data (links from other sites, press, awards, a Business Profile): without it,
the row reads "not checked: no outside data", with no stars, and the overall line names the
rows it averaged. If analytics/eeat.md exists (search-console-insights), read its newest
section and say what changed. Local: Google Business Profile, the same name, address and phone
everywhere, LocalBusiness markup, the city or region in titles and headings, hreflang
where site_languages has several, a map or directions link. website-seo-geo's "GEO" is
generative engine optimization, not local SEO.
Cloudflare zone settings (the pre-launch run when the domain is already on Cloudflare; the
live run always; never the baseline): the owner reads five places in the Cloudflare dashboard,
or shows a screenshot. The assistant never signs in and never asks for a token. The places, and
what each should show, are in docs/CLOUDFLARE-CHECKLIST.md in the Croftweaver folder, section
5, "Read in the dashboard": point the owner there rather than copy the paths, which Cloudflare
moves. The report records each place as seen, dated.
- Pre-launch, domain not on Cloudflare yet: under "Not checked yet", "Cloudflare zone settings:
deferred to the live re-check: the domain is not on Cloudflare yet". It does not count in the
verdict.
- A setting that differs from the checklist (Bot Fight Mode, AI Labyrinth, the AI policies with
Personal Agents, Bot Preference Sync, a Block Crawler switch): ask whether it was the owner's
choice (§4). Yes: "owner's choice, ", with the checklist's trade-off in one line, quoted
and credited to the checklist. No or don't know: 🔴, with the checklist item that sets it right.
- Security rules: every rule on the page, in any section, listed with its name, its action, what
it matches and "this domain" or "all websites in account", and what it does in plain words,
from what the owner reads out.
- It blocks or challenges, and nobody can say what it does or whose it is, or it could stop
visitors or crawlers the owner wants: 🔴 until the owner keeps, changes or removes it.
- Its action is Allow or Skip and nobody can say whose it is: 🟡, until the owner finds out
(an account-wide rule may be another site's).
- The owner knows it and keeps it: "owner's choice, ".
- Cloudflare's own notes (Response rules' "Cloudflare default protection applied", a banner
about rules Cloudflare deployed): quoted, no finding.
- The setup tokens, live run only (never ask at pre-launch: launch day may still need one): the
owner reads both token lists and names every token made for this site's setup and launch day.
Active or Expires soon: 🔴, delete it now. Expired: 🟡, delete it. In neither list: fine. The
owner cannot tell which they were: 🟡, find out what each token they cannot place is for, and
delete it only once it proves to be this site's setup token. Name no other token
in the report. If launch day still needs the token, read the rest and list "the setup tokens:
deferred: launch day still needs it; read again when it is done" under "Not checked yet", with
no finding.
- Not read (no dashboard access, or the owner stops): "not checked: ", 🟡, never ✅.
4. Questions and findings
Ask only where the owner's intent decides; everything else is a finding with an offer to fix.
- Ask: "Your settings say testimonials. The site has none. Still planned, or skip?" Intent
decides.
- Do not ask: a page with two H1s, a missing description, an unfilled imprint line. Those are
findings: "Found: … Want me to fix it?"
Offer to fix each gap on the spot (draft a description, describe an image, fill an Impressum
line from what the owner tells you), and run the tests again after a fix.
5. The report
Write docs/launch/launch-check-<YYYY-MM-DD>-<baseline|prelaunch|live>.md from
references/report-template.md, in the owner's language, without the template's comment. Each
run starts from the template and this run's findings, never from an earlier report.
A second report of the same kind on one day gets -2 (then -3), and so does a second
comparison: never overwrite a report.
- Verdict: 🔴 when any blocker is open; 🟡 when only soon-after items are; 🟢 when nothing
is. The baseline's verdict line is "Baseline, not a launch verdict".
- Every row says what was seen and where. A row with no evidence reads "not checked" with the
reason; never a guessed score or number.
Then make the HTML page and keep every page current. The name in braces stands for the folder
this SKILL.md is in: Claude Code fills it in; with another assistant, put that folder's path in
its place. Run the commands in the site's folder (Claude Code asks before each, since the
site's permission list does not include python3):
python3 "${CLAUDE_SKILL_DIR}/scripts/report_html.py" docs/launch/<report>.md --lang <report language>
python3 "${CLAUDE_SKILL_DIR}/scripts/report_html.py" --rebuild docs/launch
The first writes <report>.html beside the Markdown; the second rebuilds every HTML page in
docs/launch/ that has its Markdown beside it, so a report someone edited since shows the edit
(a Markdown with no page yet needs the first command). The
Markdown is the source; never edit the HTML by hand. The page opens from disk in any browser
and loads nothing.
6. Todos (when the owner wants them)
Ask: "Want to assign these?" If yes, write docs/launch/launch-todos-<YYYY-MM-DD>.md from
references/todos-template.md (Markdown only: people tick its boxes), with the names or roles
the owner gives. Anything assigned to the assistant: offer to do it now. If the project uses a
task tracker, offer to create the tickets there too.
7. Relaunch: the current site against the new one
When the settings name replaces, the pre-launch run also writes
docs/launch/launch-comparison-<YYYY-MM-DD>.md and its HTML, following references/relaunch.md
and references/comparison-template.md: a verdict, before-launch items, speed old and new,
E-E-A-T for both, open questions, a redirect map for every old address, a side by side, what
the old site gets wrong, what the new one adds and what it does not have yet.
8. The live re-check (12a)
Only once the new site's production build is Active on the live domain: never request a page
of the new site on the live domain before that (PUBLISHING.md, "Never request a page on the
live domain before its build is Active": a first request caches a 404). Reading the old site's
existing addresses before the switch is fine. Then check HTTPS, the 404 status, console errors
on the live pages and the consent question, measure the live domain's speed as website-qa §2
does (the assistant runs Lighthouse; its fallbacks are PageSpeed Insights, then the owner), and
fill the comparison's new-site row. Then read the Cloudflare zone settings (§3). The speed and
comparison parts may be skipped; the zone read may not, since for a new domain nothing else
reads it. For a WordPress move,
point at wordpress-inspector's scripts/check_move.py --moves where that skill is installed;
a site's bundled skills do not include it, so otherwise say it comes from Croftweaver.
The analytics tag on the live page. Run, in the site's folder, with the live address:
python3 "${CLAUDE_SKILL_DIR}/scripts/check_live_analytics.py" https://<live address>
It fetches the home page and compares its analytics tag with src/config.ts. A ✗ (exit 1) is a
🔴; quote its lines. Tag missing: the page carries no analytics script tag. Tag present with analytics off:
visits may be counted without the privacy page saying so. Only the script's path or id seen: 🔴
until someone looks. Exit 2 is 🟡 "not checked: ". It reads this folder's
src/config.ts, so that file must be the live one: after git fetch, run git diff --quiet origin/<PROD_BRANCH> -- src/config.ts in the site's folder (<PROD_BRANCH> is the value of
PROD_BRANCH in that file). If it fails, the row is 🟡 "not checked:
src/config.ts here differs from the live branch". Baseline and pre-launch runs: "not checked
yet: not deployed".
Right after a publish, the live page may still come from a cache, so its answer can be about the
build before. Run it again a minute later before acting on a ✗. Measured 2026-11-09 with
curl -I: croftweaver.com (Pages) answers cf-cache-status: DYNAMIC for / and /?cb=<random>;
guide.croftweaver.com (Workers static assets) answers HIT for both, so on Workers a fresh query
is no proof of a fresh page.
9. Committing
The reports are the site's records: commit them the way the site's AGENTS.md says (a branch
and a pull request, unless it says otherwise). Both the Markdown and the HTML go in. In a
public site repository, ask before committing a comparison (it lists the old site's
addresses) or todos that name people.
Related skills
website-qa: the site's test suite, which this check runs first, and the speed measurement
(§2).
website-review: the two-pass review at the end of a build.
website-testimonials: the consent rule for named quotes.
wordpress-inspector: the inventory, the redirects and the move check for a WordPress
relaunch.
outgoing-link-audit and internal-link-audit: link liveness and orphan pages.
seo-audit and website-seo-geo: deeper SEO and the AI-search contract.
search-console-insights: the analytics/ folder, including eeat.md.
1---2name: website-launch-check3description: >-4 Launch readiness check for a site built with this toolkit: runs the site's own build and tests5 first, then checks only what they cannot (legal pages, E-E-A-T and local signals, content the6 site's settings call for, testimonial consent, what only a deployed site shows, the Cloudflare7 zone settings the owner reads in the dashboard), asks only where the owner's intent decides,8 and writes a dated report with a verdict to docs/launch/ as Markdown plus a self-contained9 HTML page. When the site replaces an older one, also a current-vs-new comparison with a10 redirect map for every old address. Runs three times in new-website (baseline after the11 scaffold, the verdict before launch, the live re-check once launch day's work is done) and on12 its own. Trigger phrases: "launch check", "ready to launch", "what's missing before launch",13 "pre-launch", "go live", "launch audit", "relaunch", "compare the old and the new site",14 "post-launch gap analysis".15metadata:16 source: https://skillmd.com/skills/karero/website-launch-check17---1819# Website launch check2021A conversation, not a static report: settings first, then the site's own tests, then the22checks the tests cannot do, a question only where the owner's intent decides, an offer to fix23each gap, and a dated report. The skill's text is English; everything the owner sees is in the24owner's language (§1).2526When it runs:2728| Run | When | Report |29|---|---|---|30| **baseline** | `new-website` step 5a: after the scaffold is green (and after the brand-kit intake, where there is one) | `launch-check-<date>-baseline.md`, verdict "Baseline, not a launch verdict" |31| **pre-launch** | `new-website` step 10c: after steps 7 to 10b, before §4's deploy | `launch-check-<date>-prelaunch.md`, verdict 🟢 / 🟡 / 🔴; §4 starts after 🟢, or a 🟡 the owner accepts |32| **live** | `new-website` step 12a: once launch day's work is done (the production build Active on the live domain, the domain attached); not optional, since the zone read (§3) runs only here for a new domain | `launch-check-<date>-live.md` |3334It also runs on its own, on any site built with this toolkit, whenever the owner asks.3536## 0. Settings: `docs/launch/config.json`3738Read `docs/launch/config.json`. If the site has an older `.launch-check-config.json` in its39root, move it there first: `git mv` when git tracks it, `mv` when not, and say so. If both40exist, move nothing and ask which to keep. Make the folder first (`mkdir -p docs/launch`). Older reports in `docs/` stay where they are. The41old file's keys `storybrand` and `contact_forms` become `story_layer` and `contact_form`; fill a42missing key as below. A moved file counts as saved settings (the last paragraph of this section).4344Without a settings file, fill one from what the site already says, and ask only what is45still open:4647| Setting | From |48|---|---|49| `business_in_germany` (where the business is based decides the legal checks, not the site's language) | ask; the interview does not ask it |50| `site_languages` | `SITE.locale` in `src/config.ts`, the README's interview answers (Q4) |51| `story_layer` | the README's record of the §2a story offer, `STORY.md`; never ask again |52| `blog`, `contact_form` | the README's interview answers (Q1, Q2), the pages |53| `testimonials` | `src/data/testimonials.ts` (from `website-testimonials`); ask if the site has none |54| `analytics` | `USE_ANALYTICS` and the README (Q5) |55| `report_language` | the owner's language (§1) |56| `replaces` | ask once: does this site replace one that is live now? Its address, or `null` |5758```json59{60 "business_in_germany": true,61 "site_languages": ["de"],62 "story_layer": false,63 "blog": false,64 "testimonials": true,65 "contact_form": true,66 "analytics": "plausible",67 "report_language": "de",68 "replaces": "https://www.example.com"69}70```7172On later runs, name the saved settings in one line and ask once whether any changed.7374## 1. Language7576Questions, findings, reports, todos and the HTML are in the language the owner writes in;77`report_language` is the default on a re-run, and the owner's message overrides it. A German78owner gets everything in German, headings included. This file stays English.7980## 2. The site's tests first8182Run, in the site's folder, `npm run build`, then the whole suite with83`npx playwright test --forbid-only`, keeping the whole output. Use it, not `npm test`, so a84narrowed `test` script cannot hide a failure, and `--forbid-only` so a `test.only` left in a spec85fails the run instead of narrowing it; if `package.json`'s `test` script is anything but `playwright test`, that is86a 🟡 finding. Then follow `references/test-coverage.md`:8788- A failing test is a 🔴 blocker that names the test and quotes its message. No look by eye89 stands in for a test.90- Every `⚠ placeholders:` line is a blocker (a page still listed in `UNFILLED_UNTIL_LAUNCH`,91 or a `[MISSING: …]` placeholder, which only warns outside CI). `⚠ positioning coverage:` is a92 🟡 finding. A skipped test is a 🟡 finding.93- A category the mapping gives to a test is left to that test, **per page**, only where the94 output shows that page's test passed. The report lists those under "Left to the site's tests,95 as they stand in this site, for the pages in `PAGES`", and the check does not look at them96 again on those pages. Every other page, and every built page that `PAGES` lacks, is the97 check's.9899## 3. What the launch check checks100101Only what the tests cannot. Skip what the settings rule out, and list it under "Skipped by the102settings".103104**Always:**105- Titles and descriptions: no two pages share one, and each says what its page is (🟡).106- What no test reads (🔴, since a visitor would read it): the starter's example values (the name107 `Example`, `hello@example.com`) on the home page and the legal pages (the imprint where the108 site has one, the privacy or Datenschutz page), `public/llms.txt`'s own109 `[bracket]` slots, the 404 page's text.110- Console errors on the local preview (`npm run preview`), read with the site's own Playwright (a111 short script that opens each built page and collects console errors). Run it after the tests,112 and stop it when done: the tests start a preview of their own, and Astro refuses it while113 another one runs. Errors that come from114 the `_headers` rules, HTTPS and the 404 status show only on a deployed site: at the115 baseline and pre-launch runs they read "not checked yet: not deployed", and the live run116 checks them.117- The legal pages' answer (`README.md` in the folder that holds `package.json`): quote, in the118 report's Imprint / privacy row, every line in one of the three forms `AGENTS.md` §5 gives119 (`legal pages: reviewed by a lawyer (<date>)`, `legal pages: own texts (<date>)`, `legal pages:120 published as drafted (owner's decision, <date>)`), and ask the owner about them at every121 pre-launch run, as below.122 - No such line: a blocker (🔴): the owner's answer is not recorded; `AGENTS.md` §5 asks for it.123 Do not ask here. Claim no test failure: the site's tests pass a legal page still listed in124 `UNFILLED_UNTIL_LAUNCH` whose draft notice shows.125 - Which pages each line covers: the README names none. With one line, it covers every legal126 page. With more than one, ask the owner at each pre-launch run which legal pages each line127 was recorded for. A page's answer is the newest line that covers it (of two with one date,128 the one further down the README); quote older lines as history, with no question.129 - A page whose answer is `reviewed by a lawyer` or `own texts`: at each pre-launch run, quote130 that line and ask the owner whether it is still true for the pages it is the answer for, one131 question per line. "Yes" goes in the report as "owner confirmed, <date>". "No" or "don't132 know" is a blocker that points at `AGENTS.md` §5: record a true answer, then launch.133 - A page whose answer is `published as drafted`: no question. Report it as 🟡, one per such134 line, naming the pages it is the answer for and leaving out any the date check below135 blocks, and none for a line left with no pages: "<pages>: the kit's drafts, published on136 your decision of <date>; a lawyer's review is still open to you."137 - A legal page no line covers, or added after its line: at every pre-launch run, for each138 legal page in `tests/_helpers.ts` `PAGES` (last path segment `privacy`, `datenschutz`,139 `legal-notice` or `impressum`), run140 `git log --diff-filter=A --format=%cs -- <its file under src/pages/>`, one file per call,141 and take the first line: the date it last entered git under that name. Not `--follow`,142 which dates a renamed file (the `/datenschutz` or `/impressum` swap) by the file it came143 from. A page rewritten under the same file name keeps its first date, so this check cannot144 see the rewrite. A page that no line covers, or whose date is after its answer's date, is a blocker145 naming that page: `AGENTS.md` §5 asks again for it and adds a dated line. The answer's own146 day counts as covered. No date (no output: the file is not committed; or an error: not a147 git repo) is a blocker of its own: commit the page, then run the check again. This holds148 whatever the answer, `published as drafted` included. Claim no test failure: the tests149 compare no dates.150151**Business in Germany:**152- **Legal Notice or Impressum** per § 5 DDG (the Digitale-Dienste-Gesetz replaced the TMG in May 2024).153 The kit ships the English page at `/legal-notice`; the German one at `/impressum` replaces it on a154 German-language site, and an owner based in Germany may route both. Either address counts as the155 imprint. It carries: name,156 address, a fast electronic contact, register entry, VAT ID if any; for a regulated157 profession (architects, lawyers, doctors, and a trade that must register with a chamber such158 as the Handwerkskammer) also the chamber, the professional title and the state that granted it.159- **Datenschutz:** hosting, analytics, forms, any embeds, the contact for data questions.160 Whether the text fits what the site really does; the slots and the analytics settings are161 `new-website` §4's own item.162- **Consent banner:** needed only if the site stores or reads non-essential data on the163 visitor's device (§ 25 TDDDG, formerly TTDSG), such as Google Analytics, YouTube or Google164 Maps embeds. Analytics that store and read nothing on the device need no § 25 consent (the165 privacy page still names their legal basis, which is the owner's to settle), but check it on166 the deployed site (the browser's storage panel after a page view) rather than167 assume it from the product's name. If anything is stored or read, or you cannot tell, report168 the banner question as open for the owner, not as passed. A banner the site does not need169 is itself a finding.170171**Business elsewhere:** a privacy policy (GDPR-ready where the site serves people in the EU),172and terms where the site sells or takes bookings.173174**Story layer:** the home page against `STORY.md` (from `website-story`), not against a175generic list.176177**Blog:** an index page, at least one post (or drafts ready), a feed, a named author.178179**Testimonials** (`src/data/testimonials.ts`): real names and roles where the owner chose names, and **consent on file for180each named person**: at each pre-launch run, ask the owner, quote by quote. "Yes" goes in the report as "owner confirmed181consent on file, <date>". "No" or "don't know" for a quote is a blocker that points at182`website-testimonials`' consent rule (keep it off the live site until consent lands). Keep no183consent record and no contact details.184185**Contact form:** it shows its confirmation; a privacy note beside it links the privacy page;186spam protection where it sends to a server. Ask before sending a test message: on a live site187it reaches a real inbox.188189**E-E-A-T and local signals** (an assessment, not a measurement): rate Experience, Expertise,190Authority, Trust and Local from what the site shows, 1 to 5 with one line why. **Authority191needs outside data** (links from other sites, press, awards, a Business Profile): without it,192the row reads "not checked: no outside data", with no stars, and the overall line names the193rows it averaged. If `analytics/eeat.md` exists (`search-console-insights`), read its newest194section and say what changed. Local: Google Business Profile, the same name, address and phone195everywhere, `LocalBusiness` markup, the city or region in titles and headings, `hreflang`196where `site_languages` has several, a map or directions link. `website-seo-geo`'s "GEO" is197generative engine optimization, not local SEO.198199**Cloudflare zone settings** (the pre-launch run when the domain is already on Cloudflare; the200live run always; never the baseline): the owner reads five places in the Cloudflare dashboard,201or shows a screenshot. The assistant never signs in and never asks for a token. The places, and202what each should show, are in `docs/CLOUDFLARE-CHECKLIST.md` in the Croftweaver folder, section2035, "Read in the dashboard": point the owner there rather than copy the paths, which Cloudflare204moves. The report records each place as seen, dated.205- Pre-launch, domain not on Cloudflare yet: under "Not checked yet", "Cloudflare zone settings:206 deferred to the live re-check: the domain is not on Cloudflare yet". It does not count in the207 verdict.208- A setting that differs from the checklist (Bot Fight Mode, AI Labyrinth, the AI policies with209 Personal Agents, Bot Preference Sync, a Block Crawler switch): ask whether it was the owner's210 choice (§4). Yes: "owner's choice, <date>", with the checklist's trade-off in one line, quoted211 and credited to the checklist. No or don't know: 🔴, with the checklist item that sets it right.212- Security rules: every rule on the page, in any section, listed with its name, its action, what213 it matches and "this domain" or "all websites in account", and what it does in plain words,214 from what the owner reads out.215 - It blocks or challenges, and nobody can say what it does or whose it is, or it could stop216 visitors or crawlers the owner wants: 🔴 until the owner keeps, changes or removes it.217 - Its action is Allow or Skip and nobody can say whose it is: 🟡, until the owner finds out218 (an account-wide rule may be another site's).219 - The owner knows it and keeps it: "owner's choice, <date>".220 - Cloudflare's own notes (Response rules' "Cloudflare default protection applied", a banner221 about rules Cloudflare deployed): quoted, no finding.222- The setup tokens, live run only (never ask at pre-launch: launch day may still need one): the223 owner reads both token lists and names every token made for this site's setup and launch day.224 Active or Expires soon: 🔴, delete it now. Expired: 🟡, delete it. In neither list: fine. The225 owner cannot tell which they were: 🟡, find out what each token they cannot place is for, and226 delete it only once it proves to be this site's setup token. Name no other token227 in the report. If launch day still needs the token, read the rest and list "the setup tokens:228 deferred: launch day still needs it; read again when it is done" under "Not checked yet", with229 no finding.230- Not read (no dashboard access, or the owner stops): "not checked: <reason>", 🟡, never ✅.231232## 4. Questions and findings233234Ask only where the owner's intent decides; everything else is a finding with an offer to fix.235236- Ask: "Your settings say testimonials. The site has none. Still planned, or skip?" Intent237 decides.238- Do not ask: a page with two H1s, a missing description, an unfilled imprint line. Those are239 findings: "Found: … Want me to fix it?"240241Offer to fix each gap on the spot (draft a description, describe an image, fill an Impressum242line from what the owner tells you), and run the tests again after a fix.243244## 5. The report245246Write `docs/launch/launch-check-<YYYY-MM-DD>-<baseline|prelaunch|live>.md` from247`references/report-template.md`, in the owner's language, without the template's comment. Each248run starts from the template and this run's findings, never from an earlier report.249A second report of the same kind on one day gets `-2` (then `-3`), and so does a second250comparison: never overwrite a report.251252- **Verdict:** 🔴 when any blocker is open; 🟡 when only soon-after items are; 🟢 when nothing253 is. The baseline's verdict line is "Baseline, not a launch verdict".254- Every row says what was seen and where. A row with no evidence reads "not checked" with the255 reason; never a guessed score or number.256257Then make the HTML page and keep every page current. The name in braces stands for the folder258this SKILL.md is in: Claude Code fills it in; with another assistant, put that folder's path in259its place. Run the commands in the site's folder (Claude Code asks before each, since the260site's permission list does not include `python3`):261262```bash263python3 "${CLAUDE_SKILL_DIR}/scripts/report_html.py" docs/launch/<report>.md --lang <report language>264python3 "${CLAUDE_SKILL_DIR}/scripts/report_html.py" --rebuild docs/launch265```266267The first writes `<report>.html` beside the Markdown; the second rebuilds every HTML page in268`docs/launch/` that has its Markdown beside it, so a report someone edited since shows the edit269(a Markdown with no page yet needs the first command). The270Markdown is the source; never edit the HTML by hand. The page opens from disk in any browser271and loads nothing.272273## 6. Todos (when the owner wants them)274275Ask: "Want to assign these?" If yes, write `docs/launch/launch-todos-<YYYY-MM-DD>.md` from276`references/todos-template.md` (Markdown only: people tick its boxes), with the names or roles277the owner gives. Anything assigned to the assistant: offer to do it now. If the project uses a278task tracker, offer to create the tickets there too.279280## 7. Relaunch: the current site against the new one281282When the settings name `replaces`, the pre-launch run also writes283`docs/launch/launch-comparison-<YYYY-MM-DD>.md` and its HTML, following `references/relaunch.md`284and `references/comparison-template.md`: a verdict, before-launch items, speed old and new,285E-E-A-T for both, open questions, a redirect map for every old address, a side by side, what286the old site gets wrong, what the new one adds and what it does not have yet.287288## 8. The live re-check (12a)289290Only once the new site's production build is Active on the live domain: never request a page291of the new site on the live domain before that (`PUBLISHING.md`, "Never request a page on the292live domain before its build is Active": a first request caches a 404). Reading the old site's293existing addresses before the switch is fine. Then check HTTPS, the 404 status, console errors294on the live pages and the consent question, measure the live domain's speed as `website-qa` §2295does (the assistant runs Lighthouse; its fallbacks are PageSpeed Insights, then the owner), and296fill the comparison's new-site row. Then read the Cloudflare zone settings (§3). The speed and297comparison parts may be skipped; the zone read may not, since for a new domain nothing else298reads it. For a WordPress move,299point at `wordpress-inspector`'s `scripts/check_move.py --moves` where that skill is installed;300a site's bundled skills do not include it, so otherwise say it comes from Croftweaver.301302**The analytics tag on the live page.** Run, in the site's folder, with the live address:303304```305python3 "${CLAUDE_SKILL_DIR}/scripts/check_live_analytics.py" https://<live address>306```307308It fetches the home page and compares its analytics tag with `src/config.ts`. A ✗ (exit 1) is a309🔴; quote its lines. Tag missing: the page carries no analytics script tag. Tag present with analytics off:310visits may be counted without the privacy page saying so. Only the script's path or id seen: 🔴311until someone looks. Exit 2 is 🟡 "not checked: <its reason>". It reads this folder's312`src/config.ts`, so that file must be the live one: after `git fetch`, run `git diff --quiet313origin/<PROD_BRANCH> -- src/config.ts` in the site's folder (`<PROD_BRANCH>` is the value of314`PROD_BRANCH` in that file). If it fails, the row is 🟡 "not checked:315`src/config.ts` here differs from the live branch". Baseline and pre-launch runs: "not checked316yet: not deployed".317318**Right after a publish, the live page may still come from a cache, so its answer can be about the319build before. Run it again a minute later before acting on a ✗.** Measured 2026-11-09 with320`curl -I`: croftweaver.com (Pages) answers `cf-cache-status: DYNAMIC` for `/` and `/?cb=<random>`;321guide.croftweaver.com (Workers static assets) answers `HIT` for both, so on Workers a fresh query322is no proof of a fresh page.323324## 9. Committing325326The reports are the site's records: commit them the way the site's `AGENTS.md` says (a branch327and a pull request, unless it says otherwise). Both the Markdown and the HTML go in. In a328**public** site repository, ask before committing a comparison (it lists the old site's329addresses) or todos that name people.330331## Related skills332333- `website-qa`: the site's test suite, which this check runs first, and the speed measurement334 (§2).335- `website-review`: the two-pass review at the end of a build.336- `website-testimonials`: the consent rule for named quotes.337- `wordpress-inspector`: the inventory, the redirects and the move check for a WordPress338 relaunch.339- `outgoing-link-audit` and `internal-link-audit`: link liveness and orphan pages.340- `seo-audit` and `website-seo-geo`: deeper SEO and the AI-search contract.341- `search-console-insights`: the `analytics/` folder, including `eeat.md`.