Companion to da790de. The previous wording said 'report the gap and
stop', which agents read as 'report the gap and substitute'. Name the
near-miss substitution explicitly and point at the marker instead.
The checker gave agents no legal way to be faithful. Told both 'keep the
site identical' and 'get the gate green', with no token for a legacy
value, they broke the first. Task 10 mapped 12px and 14px both to
var(--step-1) (15px). Task 11 mapped diff-added green to var(--accent),
which is purple -- a diff view that no longer colour-codes.
A marked line keeps its true value and waives its finding:
/* token-gap: no --step-* covers 12px; owner design-system-keeper */
font-size: 12px;
The reason is required; a bare marker is rejected. Marked values are
listed on every run, so this is a visible debt queue, not a mute button.
Claude Code against an Ollama-backed model via the headroom hub. Model
defaults to glm-5.3:cloud, overridable with OC_MODEL. Unproven here, so
route it at tasks whose failure is cheap to detect.
Task 07 wrote px font sizes as the `font:` shorthand in two components
because check-tokens.mjs only matches `font-size:`. Green branch, two
hardcoded values. Make the expectation explicit: report the gap, stop.
Also fixes gates.md telling agents to rebase WIP commits away, which
git-worktrees.md forbids outright.
Task 03 rebased its branch onto task 01 and flattened the merge into four
duplicate commits — but the rule file was *telling* it to: "Before you start:
git rebase origin/main". Replaced with merge-based guidance and an explicit
prohibition, since a rewritten task branch is the same divergent-history trap
that broke the pages branch.
Also corrects gates.md tier 3, which promised screenshot comparison in CI that
is not wired in and cannot be until visual-regression.mjs grows a compare mode.
HANDOVER.md now reflects 01-04 merged, pnpm, and carries forward the review
findings that were noted but deliberately not fixed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Ten git worktrees each carried their own 225 MB node_modules (1.1 GB across
five) and paid 11s per `npm ci`. pnpm hardlinks from a shared store: the same
five worktrees cost ~250 MB total, and a fresh install is 4s.
What changed beyond the mechanical rename:
- `overrides` moved to `pnpm-workspace.yaml`. pnpm 11 does not read the `pnpm`
field in package.json *or* npm's top-level `overrides`, and it fails silently
— the vite/defu/language-server pins would have quietly stopped applying.
- Build scripts are blocked by default in pnpm; esbuild and sharp are allowed
explicitly via `allowBuilds` (renamed from `onlyBuiltDependencies` in 11).
- `packageManager` + `engines` pin the toolchain.
- gate.sh rejects a package-lock.json/yarn.lock/bun.lock outright, so an agent
running `npm install` out of habit fails loudly instead of building a second,
divergent dependency tree.
- CI bootstraps pnpm with `npm install --global pnpm@11.25.0` rather than
corepack (unbundled as of Node 25) or pnpm/action-setup (this self-hosted
act-runner has never run a job; fetching a third-party action is not
something to discover on the first one).
Two pre-existing CI bugs fixed while in the file:
- the gate installed with `npm install --package-lock=false`, which discarded
the lockfile the previous session had just fixed.
- the visual-regression step imported `playwright`, which is not a dependency,
and `visual-regression.mjs` has no compare mode anyway — in CI it overwrote
its own baselines and passed unconditionally. Removed with a comment; it
comes back when it can diff.
The `publish` job is now manual (`workflow_dispatch`). During the migration
dist/ holds three HTML files against the live pages branch's ten, so publishing
on every push to main would take the site down to a stub. Restore at task 20.
HANDOVER.md's incident log still says npm where it describes what happened at
the time; that is history, not a missed rename.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Capture rendered-text and visual baselines before route migration, and detect external CSS dependencies. Do not change verify.mjs assertions; task 02 must remove the known legacy font URL before the audit can pass.
The launcher read the agent out of each brief with [a-z-]*, which cannot match
content-i18n-migrator. Tasks 04, 05 and 06 silently fell back to
astro-architect. Now also fails loudly if the name is missing or has no agent
definition, rather than launching the wrong specialist.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The pre-existing root package.json has two scripts and no dependencies, and
there is no package-lock.json until task 01 scaffolds Astro, so every launch
died on `npm ci` before reaching the agent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Routes each task to its CLI per MODEL-ROUTING.md (Codex for 01/03/15/16/19,
Gemini for 02/18, MiniMax for the rest), creates the worktree, installs the
toolchain once task 01 has produced a package.json, and runs the agent headless
with permission prompts disabled — these run unattended, and a blocked edit
would just hang.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the agent-facing workspace and a 20-task plan for migrating the site
to Astro. Nothing here implements the refactor; these are briefs, rules and
templates that the task agents read.
- .agents/ holds context, rules, checklists, skills, specialist agents,
component/page/config templates and gate scripts. It is vendor-neutral so
MiniMax, Gemini and Codex can all read it; CLAUDE.md just points at
AGENTS.md.
- .husky/ plus .lintstagedrc.json wire the three gate tiers. gate.sh locks on
the shared git-common-dir so parallel worktrees serialise, and guards the
assertion count in scripts/verify.mjs against a coverage drop.
- plans/astro-refactor/ carries the phase graph, per-task briefs and the
model-routing recommendation.
These files must be tracked before fanning out: a worktree only checks out
tracked files, so an untracked plan is invisible to every agent working in one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Marcos Paulo