Command

PePersonal JEJust Electrolyte RTRelease Timer TtTacTile CSClearSource AtAthletic

TacTile›Launch TacTile›Web›Web — Destination editing·18 in this wave

Wave

Web — Destination editing

No destination editing exists on either web app today. apps/web has zero editing (creation is also being built from scratch in a separate wave); apps/web-mobile has creation for 7 of 14 types but no re-edit path for any of them. Neither web app has an authenticated in-app destination detail/preview page distinct from the public /t/[hash] scan page either — apps/web has no /destinations route of any kind (confirmed: zero list, detail, or edit surface in app/(app)/), and apps/web-mobile's DestinationDetailPage only shows a type badge, date, and trigger-link management, not the destination's actual rendered content. Folding the detail/preview gap in here since one screen (mirroring iOS's DestinationLaunchpad: preview + connected triggers + edit entry + delete) naturally serves both.

TacTile › Launch TacTile › Web › Web — Destination editing

12/18 done6 open1 open blocker

Selected taskclear

Build destination editing on apps/web-mobile for its 7 live types

DoneOwner · Vivt_c04492ba

Unlocks 1Approve: Destination editing

Show the prompt

You are Viv, the agent for the TacTile lane of Command.

Open the lane and work there:

```bash
cd ~/Command/tactile && claude
```

Read that lane's CLAUDE.md and lane.json before you touch anything — they
carry your identity, your scope, and the hazards specific to this lane.
Read state from the hub (https://internal.tactile.to) or ask. Never guess at it.

## The task

**Build destination editing on apps/web-mobile for its 7 live types** · done

Where it sits: TacTile › Launch TacTile › Web › Web — Destination editing

## Unlocks 1

- Approve: Destination editing

## The brief on file

DONE — PR #248, merged to main as 7dbf5c08 (ancestry verified).

Premise was stale twice over, as round 3 flagged: web-mobile created ALL 14 kinds, not 7, and had zero edit flow. Scope was therefore all 14.

SHAPE DECISION. Introduced the shared interface FIRST rather than porting per-file. All 14 builders declared their own two-member `XBuilderProps`; they now extend one `BuilderProps` in features/destinations/components/builders/types.ts, with parameterised builders (url/category, action/type, payment/provider, gallery/mode, document/title) adding their own member. `edit` is optional, so an untaught builder still compiles and creates.

`BuilderEditContext` carries `kind` as well as hashId + content — VideoBuilder serves both `video` and `audio` and they PATCH different endpoints (/api/videos vs /api/audio); in edit mode there is no file to infer from and the payload does not name its kind.

SHIPPED: app/destinations/[hash]/edit/, features/destinations/pages/EditDestinationPage.tsx, lib/editableKinds.ts (the gate as data — the root vitest config maps `@/` to apps/web, so a test importing the page would resolve web-mobile's components to the desktop app's), the update+content wrappers bound in lib/destinationApi.ts, UnsavedChangesAlert extracted for reuse, and an Edit affordance on the detail page.

TWO KINDS ARE DELIBERATELY PARTIAL, both non-destructively:
· MENU edits the header only (name/tagline/accent) and OMITS `sections` from the PATCH. `sections` is a DELETE-then-rebuild replace and /content never returns section `emoji`, so round-tripping would blank every emoji. Unblocks on t_7e7ab719.
· MEDIA kinds (document, video/audio, gallery) edit metadata only — the multipart re-upload branch is unwrapped in the package. Pickers are hidden in edit mode with an explanation rather than appearing to work.

FORM carries a guard: PUT accepts fields:[] and wipes every field (t_a9ead0cc). canSave goes false with a rendered reason when an edit would empty them.

TEST: tests/unit/web-mobile-edit-coverage.test.ts derives kinds from DESTINATION_ROUTES and scans every builder for an edit branch — the second assertion guards a builder refactored back to create-only while EDITABLE still claims it works, which would CREATE A SECOND DESTINATION on save. Both halves confirmed to fail by injecting the drift.

## How to work this

- Do the work, don't just plan it. If the fix is obvious and safe, make it.
- Surface anything needing Matthew's sign-off instead of deciding it yourself.
- Short replies. Lead with the answer, then stop.
- Finish by saying what changed, what you verified, and what you'd do next.

To doIn progressDoneBlockedPick a wave on the left · tap a task to expand it

Lanes

Web — Destination editing

Build destination editing on apps/web-mobile for its 7 live types