The settings of the plugin example listed the Gradle Plugin Portal first
and filtered only the Progmasoft repository, so Gradle asked the Portal for
com.progmasoft.tazel.plugin before Progmasoft's own repository. The
example now makes https://maven.progmasoft.com/tazel exclusive for
com.progmasoft.tazel and its subgroups: Gradle resolves those there and
nowhere else, and every other plugin from the Portal.

The example builds as written: the Gradle plugin and tazel-api resolve
and jar writes the plugin JAR with its descriptor.

The Building a plugin section asked for com.progmasoft.tazel.plugin
0.4.0 from the Gradle Plugin Portal alone. The Portal does not serve it
yet, and the released version is 0.4.0.3. The example now has the
settings.gradle.kts that adds https://maven.progmasoft.com/tazel for the
group com.progmasoft.tazel, uses version 0.4.0.3, and adds Maven
Central, without which Kotlin's standard library does not resolve.

The example was built as written: the Gradle plugin and tazel-api
0.4.0.3 resolve and gradlew jar writes the plugin JAR with its
descriptor. Progmasoft/tazel carries the same change in
Documents/PLUGINS.md for the next release, so the next sync keeps it.

Name the crawlers robots.txt lets in

Add CodeQL, Codecov, Codacy and actionlint

zizmor found that the pinned commit was not the one the v4 comment
named; the pin is the commit of v4.39.3 and says so.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

- CodeQL analyses the workflows and the TypeScript and Vue sources with
  the security-extended queries, on every push to main, every pull
  request and once a week.
- The tests run under c8, and a job of its own uploads the lcov report
  to Codecov through OIDC, away from the job that installs packages.
- .codacy.yml leaves out what is generated and what is copied from
  Progmasoft/tazel and Progmasoft/tala.
- actionlint joins zizmor in Workflow Security.
- CI Required gathers the site job and Dependency Review, so a ruleset
  can require one check.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Search crawlers and the crawlers that fetch a page for a person are
allowed: Googlebot, Google-Extended, OAI-SearchBot, ChatGPT-User,
Claude-SearchBot, Claude-User and ClaudeBot. GPTBot, CCBot and
Bytespider are turned away. Every other crawler is allowed, and the
sitemap stays named.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

tazel.net is the site of Tazel itself: a Nuxt 4 site generated to static
files, in en-US, de-DE, ru-RU and he-IL, light by default with a dark
theme that is kept per browser.

- Home: the release and its kind, what Tazel does, the steps from a
  project file to a program, and a sample.
- Docs: Markdown through Nuxt Content, highlighted by Shiki (Kotlin,
  shell, PowerShell, and Tala through the grammar of the compiler
  repository), with a tree of the pages on the left, the headings of
  the page on the right, and the previous and next pages below.
  Installation, a first project, and releases and support are written
  in every language; the reference (commands, project file, lockfile,
  plugins, how Tazel works) is copied from the Documents of Tazel by
  scripts/sync-docs.mjs, and a language without a page shows the
  English one with a note.
- Every page has its canonical address, hreflang alternates and a
  sitemap entry; generation fails on a link to a page that does not
  exist.

CI checks the format, the types and the language helpers, generates
the site and runs Dependency Review on pull requests.