- auto-release.yml: the by-hand test run passes github.sha (what it got anyway) instead of an
output that also reads workflow_run.head_sha, so CI never checks out a ref a workflow_run
handed over in a job that uses caches (cache poisoning, 7 alerts in ci.yml).
- tests/harness.js: the script splitter also matches end tags like </SCRIPT > (bad HTML
filtering regexp).
- Tests paste the fixture saves into the page with lit(), which escapes what could end a
script or a string, instead of plain JSON.stringify (improper code sanitization).
A version bump on main ran the whole suite twice: once in CI for the
push, once more in auto-release. Auto-release now starts when CI
finishes on main and releases only if that run was green, tagging the
exact commit CI tested. Started by hand, it still tests first, since
there is no CI result to go by.
- The harness builds the DOM without the page's 2 MB of script text and
runs the scripts straight in the window: a page load in tests takes
about half the time (jsdom's parser was slow on the script text).
- That also took the game's source out of the DOM, which showed three
tests that matched a phrase in the page source rather than the screen;
they now check the panel, the routing window and the dictionary.
- CI plans its split once and runs the tests on 6 runners; every run on
main records each file's time and caches it, so the parts stay even
without anyone refreshing tests/durations.json.
- auto-release.yml and release.yml call the same sharded tests instead
of running all 170 files on one runner each (11 and 15 min); a release
is tested once, before it is tagged.
- Audit: five tests that duplicated a gameplay rule or checked nothing
are gone; the two slowest checks sample a third of the cities, and a
year of play to reach one due date became a moved due date.
- auto-release: pass step outputs to run scripts through env, not ${{ }} interpolation
- perf-comment: download the PR's perf artifact to RUNNER_TEMP, outside the workspace
- dev server: log errors to the terminal instead of sending stack traces in the response
- i18n-add/i18n-check: escape every regex metacharacter when building patterns from labels
The image is published to GHCR as the release's version when a release
dispatches docker.yml, and never as latest or per commit. Pushes and pull
requests that touch it build it and try headless (with and without a
token) and browser mode, without publishing.
The image: node:24-slim pinned by digest (the major CI tests on), npm ci
from the lockfile in its own layer, only src, site, scripts and the
harness copied in (an allow-list .dockerignore), root-owned and run as
node, /saves ready for --save-dir. The entrypoint runs the servers with
node directly and, with no SOT_MCP_TOKEN, makes a token for the run and
prints it (MCP_NO_TOKEN=1 to opt out); the docs publish ports to
127.0.0.1 and explain --allow-host. Labels, SBOM and provenance on push.
Every GitHub Action in every workflow is pinned to a commit SHA with its
version, and Dependabot keeps the actions, the image's base and the npm
tools current. tests/docker.test.js guards all of it. The ADR and the
github-pages skill say the image is a tool: GHCR only, never Pages.
A new auto-release workflow runs the tests, tags v<version>, publishes a
release with that version's changelog section and starts the deploy from
the tag. It skips versions that are already tagged and refuses older
versions or ones without notes.