The real-host smoke test was the migration's #1 production-only failure mode and had never run. It has now run, without taking the site down: the Astro dist was published to `pages` additively under two previously-unused paths (`_astro/` and `_verify/summary/`), so all ten live pages stayed up, then force-pushed away. Astro's base-prefixed absolute asset URLs resolve on the Pages Server — that was the actual risk, and it is now proven rather than assumed. Trailing-slash redirects match `trailingSlash: 'always'`. Also corrects two things the guide got wrong: - a `?v=$(git rev-parse --short HEAD)` cache-busting idiom. The Pages Server caches for ten minutes keyed on path, so a query string never busted it; the guide was telling operators to trust a check that could not work. A file you just deleted keeps serving 200 until the cache expires. - the claim that a push to `main` publishes. It no longer does, and must not until cutover. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
8.8 KiB
Handover — Astro refactor, session of 2026-09-05
Phase 0 is done, merged, and green. Tasks 01–04 are all in main. The project
now runs on pnpm, not npm. Nothing has been pushed: main is 23 commits
ahead of origin/main.
Where the work is
| Branch | Head | State |
|---|---|---|
main |
see git log |
01–04 merged, gate green under pnpm |
refactor/task-01-scaffold |
71021e6 |
merged |
refactor/task-02-tokens |
c9ec9e3 |
merged |
refactor/task-03-verification-net |
2e79aac |
merged |
refactor/task-04-content-schema |
b554f87 |
merged (gate verified before merge) |
All four merged with --no-ff and zero conflicts. The four worktrees in
~/Projects/af-task-0* still exist and still hold npm-era node_modules; they
are stale. ~/Projects/af-task-02/before/ still holds 29 MB of misplaced
untracked screenshots — delete or move them before removing that worktree.
What each task actually produced
01 — scaffold (Codex). Astro 5.5 with base: '/ai-for-dummies', strict TS,
BaseLayout.astro, summary.astro as the single smoke page, lint/format/husky
configs, .gitea/workflows/, rewritten publishing section in
docs/operations-guide.md, hands-on/ copied verbatim into public/. Build
passes, gate passes, hooks verified live, 42/42 assertions intact.
02 — tokens (Gemini). src/styles/tokens.css and src/styles/base.css only
— 43 lines. It did not touch the legacy CSS, deliberately: rewriting
styles.css would break verify.mjs's exact-string assertions. So the three
drifting palettes are now resolved in the new token layer, while the live site
still runs on the old values. That is the correct scope, but it means the
consolidation is not proven visually yet.
03 — verification net (Codex). .agents/scripts/visual-regression.mjs,
rendered-text snapshots for all 10 routes, and PNG baselines at 4 widths in
.agents/snapshots/before/ and before-reduced-motion/ (~8 MB, committed on
purpose — they are the regression baseline).
04 — content schema (MiniMax). src/content/config.ts with eight empty
collections and a strict localized({en, pt}) helper. No content moved; that is
tasks 05 and 06.
Things I fixed that the plan got wrong
AGENTS.mdlistedpackage-lock.jsonunder Never touch, meaning never hand-edit. Task 01 read it as never create and shipped with--package-lock=false.npm ci— which is how every worktree spun up at the time — cannot work without it. Wording corrected; lockfile committed.- The first lockfile was reconstructed from a
node_modulesinstalled without one, so its entries had noresolved/integrityandnpm cifailed withETARGET tinyglobby@0.2.17. Regenerated from a clean install. .husky/commit-msgrejectedbuild:andrevert:. Both are standard conventional types. Added.launch.shread the agent name with[a-z-]*, which cannot matchcontent-i18n-migrator. Tasks 04/05/06 silently fell back toastro-architect. Fixed to[a-z0-9-]*and made it fail loudly instead of guessing. Task 04 was relaunched under the right agent.
Things the agents got wrong, and what I did
- Task 03 went outside its brief. It rewrote
scripts/audit-ui.mjsto ban external CSS dependencies — a sound check that the pre-existing malformed@font-faceinstyles.cssviolates, so it left the gate red for every downstream task. Reverted in2e79aac; the snapshot net was kept. The check should come back once the font decision is made. - Task 03 rebased instead of branching, flattening task 01's merge into four
duplicate commits with new SHAs — the same divergent-history trap that broke
the
pagesbranch. Replayed onto the proper base; task 01 is an ancestor again. Worth adding to.agents/rules/git-worktrees.md: never rebase a task branch onto anything. - Task 02 left 29 MB of screenshots in
~/Projects/af-task-02/before/— untracked, and in the wrong place (.agents/snapshots/is the right one). Its report claims the screenshot box is unticked while the files exist. Delete them or move them; do not commit them where they are.
Still open from task 01's own report
Real-host smoke test of— done, see "Resolved: the base path serves correctly on the real host" above./ai-for-dummies/summary/- No Playwright locally, so screenshot comparison still cannot run anywhere.
visual-regression.mjsalso has no compare mode, so it is out of CI until it grows one. See.agents/rules/gates.md.
The decision that is still yours
styles.css:1 has an @font-face whose src: points at a Google Fonts
stylesheet, not a font file. Manrope and DM Mono have therefore never
rendered; the site has always been Arial and generic monospace. Self-hosting
them during the migration would silently redesign the site. Task 02 defaulted to
matching what renders today. Decide explicitly, then task 19 can restore the
external-CSS assertion. See .agents/context/design-system.md.
Next steps, in order
- Decide the font question (see below). It blocks task 19 and it is now the only thing here that needs a human.
- Fan out phase 1 and 2:
All three are unblocked now that 02 and 04 are in
.agents/scripts/launch.sh 05 content-guide --base main .agents/scripts/launch.sh 06 content-review --base main .agents/scripts/launch.sh 07 primitives --base mainmain. After 07 lands, 08-11 run four at a time. Task 07 is the routing calibration point - judge it before committing to MiniMax for the other twelve. - Clean up the four merged worktrees and branches in
~/Projects/af-task-0*.~/Projects/af-task-02/before/holds 29 MB of misplaced untracked screenshots; delete or move them first.
Resolved: the base path serves correctly on the real host
This was the #1 production-only failure mode and it is now verified against
the live Pages Server, not just pnpm run preview.
Method, which is reusable and costs no downtime: the Astro dist/ was published
to the pages branch additively — _astro/ plus a single page at
_verify/summary/, both paths previously unused, so no existing file was
touched and all ten live pages stayed up throughout. Results:
/ai-for-dummies/_verify/summary/→ 200, and it really is the Astro page (<title>AI For Dummies — Route map</title>), not a 200-ing error page./ai-for-dummies/_astro/chapters.PwjFDHFw.css→ 200text/css. This is the actual base-path proof: Astro emits base-prefixed absolute asset URLs, and the server resolves them./ai-for-dummies/_verify/summary(no trailing slash) → 307 to the slashed form, which matchestrailingSlash: 'always'.
The probe was force-pushed away immediately; pages is back at 37a1e48.
Note for whoever publishes next: the Pages Server caches for ten minutes,
keyed on path. A deleted file keeps serving 200 until the cache expires, and a
?v=<sha> query does not bust it. See docs/operations-guide.md step 6.
Carried over from the phase 0 review
Found while reviewing the four diffs, none of it blocking, none of it fixed:
public/hands-on/is a byte-identical copy ofhands-on/with no sync mechanism. Two trees will drift. Symlink, or copy as a build step.--bp-*tokens cannot work. CSS custom properties do not resolve inside@mediaqueries. The real enforcement ischeck-tokens.mjs'sALLOWED_BREAKPOINTSallowlist; the variables are decoration and will mislead whoever writes the first responsive component.tokens.csshas no font tokens at all, and the scale skips--step-2and--step-3. Downstream of the font decision.visual-regression.mjshas no compare mode. It only captures. It is out of CI for that reason — see.agents/rules/gates.md.src/content/config.tsuses the legacytype: 'data'API. Astro 5 auto-converts it to a glob loader and it works; prefer explicitfile()/glob()loaders insrc/content.config.tswhen 05/06 touch it.
Running an agent
.agents/scripts/launch.sh <nn> <slug> [--base ref] [--cli codex|agy|mm] [--fg]
Routing is automatic: Codex for 01/03/15/16/19, agy (Gemini 3.1 Pro) for
02/18, mm (Claude Code against MiniMax-M3) for the rest. All three launch with
permission prompts disabled, because a blocked edit in an unattended run just
hangs. Logs land in .agents/logs/ (gitignored).
Task 07 is the routing calibration point. It is small and easy to judge. Check it before committing to MiniMax for the other twelve.