# Postmark Transactional Email

## Objective

Send transactional email through Postmark on the right stream, with bounces tracked.

A Postmark delivery path that separates message streams, honours suppression, and retains provider identifiers for diagnostics.

## Before You Begin

This feature is being added to an application that already exists and already
works. Do not scaffold a new project, and do not assume a blank slate.

Inspect the codebase first and establish:

- The existing application structure and where code of this kind already lives.
- The framework and version in use.
- The existing design system — colours, spacing, typography, and component conventions.
- Existing UI components you can reuse instead of writing new ones.
- The existing database structure, if this feature needs to persist anything.
- The existing authentication and authorization system, if this feature is user-scoped.
- Dependencies already installed, so you don't add a library that duplicates one.
- The existing test setup and conventions.

Only start writing code once you understand the above. If the application
already implements part of this feature, extend it rather than replacing it.

## Implementation Instructions

1. Put Postmark behind the app's existing mail layer as one adapter, so no calling code constructs a provider-specific payload. This entry is an alternative to the other transactional email integrations, not an addition — pick one adapter and keep the mail layer identical.
2. Classify every outbound message as transactional or broadcast and bind that classification to the matching message stream at the mail layer, so the choice is made once and cannot be got wrong at an individual send site.
3. Send from the app's background-job system and persist the provider's message identifier on the record the moment it comes back, so a support question about one email can be answered without guessing.
4. Consume delivery and bounce feedback on a verified endpoint and maintain a suppression state per address that every send checks before queueing.
5. Do not send transactional mail to an address the provider has marked inactive just because the app still has it on a user record. Reconcile the app's view with the provider's suppression state and surface the problem to the user instead.

## UI and UX Requirements

Match the application's existing design system exactly. Reuse its components,
spacing, and typography. This feature should look like it was always there.

## Responsive Requirements

Works on mobile, tablet, and desktop. Touch targets are large enough to hit on a
phone, and nothing overflows horizontally at 320px.

## Accessibility Requirements

- Fully keyboard navigable.
- Correct semantic elements and ARIA roles.
- Visible focus states.
- Meets WCAG AA contrast.
- Dynamic changes are announced to screen readers.
- Respects prefers-reduced-motion.

## Edge Cases

- Sending a broadcast on the transactional stream, or the reverse, degrades deliverability for everything else. Bind the stream to the message classification centrally rather than passing it in at each call site.
- An address the provider has deactivated after repeated bounces or a complaint will silently stop receiving mail. Detect the inactive state, show the affected user that their address is not receiving mail, and give an operator a way to reactivate after the address is corrected.
- Delivery and bounce webhooks arrive on a public endpoint and can be replayed. Verify their authenticity before trusting them, and apply them idempotently so a repeated event does not double-count a bounce or overwrite newer state.
- A job retried after a timeout may resend an email that already went out. Key each send on a stable identifier so the retry reconciles rather than duplicating.
- Provider message identifiers are the only reliable way to trace a specific email in a support conversation. Store them alongside the message record and keep them for as long as support needs them.
- Transactional mail must not be gated on marketing consent. A receipt or a password reset goes to someone who has unsubscribed from everything else.
- Non-production environments must not reach real recipients. Suppress delivery or redirect every message to a fixed internal address by default outside production.
- When the provider is degraded, the triggering action must still succeed, the message must stay queued for retry, and no log line should carry the credential or the full body.

## Testing

Exercise the feature end to end in the running application. Cover every edge case
above, then run the existing test suite and confirm nothing regressed.

## Acceptance Criteria

- [ ] Every message is classified and sent on the stream matching its classification, decided in one place.
- [ ] Provider message identifiers are stored on every sent message record.
- [ ] Bounce and delivery webhooks are authenticated and idempotent.
- [ ] Suppressed or inactive addresses are skipped before queueing and surfaced to the user.
- [ ] Retried jobs never produce a duplicate email.
- [ ] Non-production environments cannot deliver to real recipients.
- [ ] The feature matches the existing design system.
- [ ] No existing functionality is broken.

## Adaptation Rules

- Match the existing design system. Do not introduce a new colour palette,
  spacing scale, or component library.
- Reuse existing components and utilities wherever they fit.
- Follow the naming, file layout, and code style already present.
- Do not upgrade, replace, or remove existing dependencies to make this
  feature fit. Adapt the feature to the app, not the app to the feature.
- Do not break existing functionality. If a change is genuinely required in
  existing code, make the smallest one that works and say so.
- If something in these instructions conflicts with how the application is
  built, follow the application and explain the deviation.

## Final Verification

Before you report the work as done:

1. Re-read the acceptance criteria above and check each one against what you
   actually built.
2. Run the application and exercise the feature end to end.
3. Run the existing test suite and confirm you have broken nothing.
4. Check the feature on mobile, tablet, and desktop widths.
5. Check keyboard navigation and focus handling.
6. Summarize what changed: files added, files modified, and anything you
   deliberately did differently because of how this application is built.

If any acceptance criterion is unmet, fix it before reporting completion.