fix: deploy canonical product router

[steipete]

Summary

• source-control the existing crabfleet-canonical-router Worker
• proxy crabfleet.ai to the generic Crabbox product origin without forwarding credentials
• redirect product aliases to the product apex with path/query preserved
• deploy the product router before the app Worker in CI and manual aggregate deploys

Production finding

After PR #17 deployed, external verification showed the separately managed product router still redirected crabfleet.ai to the old app host. This makes product routing part of the repository deployment instead of relying on stale out-of-band Worker source.

Verification

• pnpm check
• pnpm test (45 tests)
• go test ./... && go vet ./...
• pnpm deploy:product --dry-run
• pnpm exec wrangler deploy --dry-run
• git diff --check
• autoreview clean: no accepted/actionable findings

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f7496afeec

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Source-control the product alias route

With workers_dev disabled and no routes/custom domains in this product config, the new deploy only uploads the crabfleet-canonical-router script; it does not make the added crabfleet.app alias testable in production. I checked the domain convergence script and it only iterates crabfleet.ai and www.crabfleet.ai, so requests to crabfleet.app will still depend on an out-of-band Cloudflare route and can drift or be absent even after this workflow succeeds.

Useful? React with 👍 / 👎.