@quickback-dev/cli 0.60.6 → 0.60.7

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/index.js CHANGED
@@ -10,7 +10,7 @@
10
10
  */
11
11
  import pc from "picocolors";
12
12
  // Version injected at build time by scripts/inject-version.ts
13
- const CLI_VERSION = "0.60.6";
13
+ const CLI_VERSION = "0.60.7";
14
14
  function getPackageVersion() {
15
15
  return CLI_VERSION;
16
16
  }
@@ -108,16 +108,16 @@ under the same `openapi.publish` gate. See [Output
108
108
  structure](/tooling/output#spec-blobs-are-assets-not-bundle).
109
109
 
110
110
  > **BREAKING — `openapi.json` is no longer written to the project root.**
111
- > Tooling that read it (`npx openapi-typescript openapi.json`, CI steps, lint
112
- > rules) must point at `src/apps/__quickback/spec/openapi.json`.
111
+ > Tooling that read it (`npx openapi-typescript openapi.json`, CI steps, lint
112
+ > rules) must point at `src/apps/__quickback/spec/openapi.json`.
113
113
  >
114
- > `api-types.gen.ts` is **unaffected** — the compiler still generates it at the
115
- > project root, which is why most projects never invoke `openapi-typescript`
116
- > directly and need no change at all.
114
+ > `api-types.gen.ts` is **unaffected** — the compiler still generates it at the
115
+ > project root, which is why most projects never invoke `openapi-typescript`
116
+ > directly and need no change at all.
117
117
  >
118
- > On your first compile after upgrading, the CLI deletes the now-orphaned root
119
- > copies. It only does so for blobs it just re-emitted under the new path, so an
120
- > older compiler that still writes to the root is never second-guessed.
118
+ > On your first compile after upgrading, the CLI deletes the now-orphaned root
119
+ > copies. It only does so for blobs it just re-emitted under the new path, so an
120
+ > older compiler that still writes to the root is never second-guessed.
121
121
 
122
122
 
123
123
  ## v0.59.1 — July 29, 2026
@@ -131,34 +131,34 @@ Null-safe scope semantics make this safe (these apply **inside `all`/`any` group
131
131
  - **Owner** (`ctx.userId`) and **soft-delete** (`isNull deletedAt`) are always present / always AND'd — never relaxed.
132
132
 
133
133
  > **`ctx.activeTeamId` is implicitly optional only while another tenant arm
134
- > bounds the query.** Dropping the team filter is safe when an org (or owner,
135
- > `via:`, `permission`, or a binding group) arm still constrains the rows —
136
- > that is what makes it "the teamless rest-of-org slice".
134
+ > bounds the query.** Dropping the team filter is safe when an org (or owner,
135
+ > `via:`, `permission`, or a binding group) arm still constrains the rows —
136
+ > that is what makes it "the teamless rest-of-org slice".
137
137
  >
138
- > When **team is the only tenant arm**, there is no rest-of-org slice to fall
139
- > back to, so an absent claim denies with ``sql`1 = 0` `` instead of dropping.
140
- > Otherwise every predicate would evaluate to `undefined`, the firewall would
141
- > return no `WHERE` at all, and any authenticated caller would read every
142
- > tenant's rows.
138
+ > When **team is the only tenant arm**, there is no rest-of-org slice to fall
139
+ > back to, so an absent claim denies with ``sql`1 = 0` `` instead of dropping.
140
+ > Otherwise every predicate would evaluate to `undefined`, the firewall would
141
+ > return no `WHERE` at all, and any authenticated caller would read every
142
+ > tenant's rows.
143
143
  >
144
- > This matters because `ctx.activeTeamId` is absent in ordinary
145
- > configurations: the Better Auth teams plugin is opt-in, and the external auth
146
- > provider does not supply the claim at all. Note that soft-delete and literal
147
- > predicates (`kind = 'public'`) are **row filters, not tenant boundaries** —
148
- > they do not count as a co-existing arm.
144
+ > This matters because `ctx.activeTeamId` is absent in ordinary
145
+ > configurations: the Better Auth teams plugin is opt-in, and the external auth
146
+ > provider does not supply the claim at all. Note that soft-delete and literal
147
+ > predicates (`kind = 'public'`) are **row filters, not tenant boundaries** —
148
+ > they do not count as a co-existing arm.
149
149
  >
150
- > A table with a `teamId` column and no `organizationId` gets team-only
151
- > isolation from auto-detection, without any explicit `firewall` block.
150
+ > A table with a `teamId` column and no `organizationId` gets team-only
151
+ > isolation from auto-detection, without any explicit `firewall` block.
152
152
  >
153
- > **You will usually see this as a compile error rather than a silent deny.**
154
- > When the teams sub-plugin is off, `ctx.activeTeamId` can never be populated,
155
- > so a team-only firewall would deny *every* request — the compiler rejects
156
- > that configuration up front and tells you to either enable teams
157
- > (`auth: { plugins: { organization: { teams: { enabled: true } } } }`, which
158
- > requires object-form plugin config) or give the resource a bindable tenant
159
- > scope. The runtime deny remains the backstop for the case the compiler can't
160
- > see, such as teams being enabled while a particular caller has not selected
161
- > a team.
153
+ > **You will usually see this as a compile error rather than a silent deny.**
154
+ > When the teams sub-plugin is off, `ctx.activeTeamId` can never be populated,
155
+ > so a team-only firewall would deny *every* request — the compiler rejects
156
+ > that configuration up front and tells you to either enable teams
157
+ > (`auth: { plugins: { organization: { teams: { enabled: true } } } }`, which
158
+ > requires object-form plugin config) or give the resource a bindable tenant
159
+ > scope. The runtime deny remains the backstop for the case the compiler can't
160
+ > see, such as teams being enabled while a particular caller has not selected
161
+ > a team.
162
162
 
163
163
 
164
164
  `nullable: true` is a separate database-schema opt-in, supported only on an
@@ -222,17 +222,17 @@ changeset writes; without it the compiler warns that changeset writes to a
222
222
  A table that declares no `read.access` is authenticated-only, identical to
223
223
  its CRUD read.
224
224
 
225
- > **This differs from `?include=` on purpose.** FK-graph `?include=`
226
- > *requires* its target to declare `read.access` and is a compile error
227
- > otherwise; a view surface does not.
225
+ > **This differs from `?include=` on purpose.** FK-graph `?include=`
226
+ > *requires* its target to declare `read.access` and is a compile error
227
+ > otherwise; a view surface does not.
228
228
  >
229
- > Neither surface is weaker than the other. A table with no declared
230
- > `read.access` is already authenticated-only-plus-firewall when read
231
- > directly through CRUD, so embedding it in a view grants nothing extra — the
232
- > view is exactly as strong as reading the table itself, which is the rule
233
- > both surfaces are held to. The difference is authoring strictness, not
234
- > enforcement: `?include=` makes you state the target's read policy before
235
- > embedding it, while a view inherits whatever the table already does.
229
+ > Neither surface is weaker than the other. A table with no declared
230
+ > `read.access` is already authenticated-only-plus-firewall when read
231
+ > directly through CRUD, so embedding it in a view grants nothing extra — the
232
+ > view is exactly as strong as reading the table itself, which is the rule
233
+ > both surfaces are held to. The difference is authoring strictness, not
234
+ > enforcement: `?include=` makes you state the target's read policy before
235
+ > embedding it, while a view inherits whatever the table already does.
236
236
 
237
237
 
238
238
  {/* doc-compile: skip — illustrative fragment: the point is the `read.access` line and the three 403s below it, so the host `customers` table and its columns are deliberately not repeated. Whole-file table examples live in /define/schema. */}
@@ -146,35 +146,35 @@ POST /storage/v1/bucket
146
146
  | `user` | Owner only | Owner only |
147
147
 
148
148
  > **`writeScope: "user"` is enforced per object, not per organization.** Any
149
- > member of the org may upload to such a bucket, so the object key is the only
150
- > thing separating one member's files from another's — and keys are built from a
151
- > client-supplied path. Before any write, Quickback checks whether the target key
152
- > already has an owner: writing to a key owned by another user returns `403`
153
- > `ACCESS_OWNERSHIP_REQUIRED`.
149
+ > member of the org may upload to such a bucket, so the object key is the only
150
+ > thing separating one member's files from another's — and keys are built from a
151
+ > client-supplied path. Before any write, Quickback checks whether the target key
152
+ > already has an owner: writing to a key owned by another user returns `403`
153
+ > `ACCESS_OWNERSHIP_REQUIRED`.
154
154
  >
155
- > This applies to presigned uploads too. A signed `PUT` URL *is* a write
156
- > capability, so ownership is checked **before** the URL is issued, not when the
157
- > bytes land.
155
+ > This applies to presigned uploads too. A signed `PUT` URL *is* a write
156
+ > capability, so ownership is checked **before** the URL is issued, not when the
157
+ > bytes land.
158
158
 
159
159
 
160
160
  > **The read policy covers metadata, not just bytes.** `readScope` and
161
- > `readRoles` are applied on every route that can reveal an object — downloads,
162
- > `GET /object` listings, and `POST /url` lookups alike. Being a member of the
163
- > owning organization is not sufficient on its own.
161
+ > `readRoles` are applied on every route that can reveal an object — downloads,
162
+ > `GET /object` listings, and `POST /url` lookups alike. Being a member of the
163
+ > owning organization is not sufficient on its own.
164
164
  >
165
- > - **Listing** (`GET /object`) returns only objects in buckets you may read.
166
- > The filter is applied in the query, so `limit` and `offset` page over your
167
- > readable set — you never receive a short page because rows were removed
168
- > after the fact. Naming a bucket you cannot read returns `403`
169
- > `ACCESS_ROLE_REQUIRED`; with no `bucket` filter, unreadable buckets are
170
- > simply absent.
171
- > - **URL lookup** (`POST /url`) returns `404` when the bucket policy denies
172
- > you — the same response as an object that does not exist. This is
173
- > deliberate: a `403` would confirm that the id or key you supplied resolves
174
- > to a real object, letting a caller enumerate objects they cannot read.
165
+ > - **Listing** (`GET /object`) returns only objects in buckets you may read.
166
+ > The filter is applied in the query, so `limit` and `offset` page over your
167
+ > readable set — you never receive a short page because rows were removed
168
+ > after the fact. Naming a bucket you cannot read returns `403`
169
+ > `ACCESS_ROLE_REQUIRED`; with no `bucket` filter, unreadable buckets are
170
+ > simply absent.
171
+ > - **URL lookup** (`POST /url`) returns `404` when the bucket policy denies
172
+ > you — the same response as an object that does not exist. This is
173
+ > deliberate: a `403` would confirm that the id or key you supplied resolves
174
+ > to a real object, letting a caller enumerate objects they cannot read.
175
175
  >
176
- > For `readScope: "user"` buckets, both routes additionally require that you own
177
- > the object.
176
+ > For `readScope: "user"` buckets, both routes additionally require that you own
177
+ > the object.
178
178
 
179
179
 
180
180
  ### Role-Based Access
@@ -334,11 +334,11 @@ The files worker:
334
334
  4. Serves the file or returns 403
335
335
 
336
336
  > Step 2 is a live lookup, not a read of the session row. A session outlives
337
- > membership changes, so its stored active organization only counts while a
338
- > current member row backs it. Remove a member and their org-scoped private-file
339
- > access stops on the next request — including on the JWT fast path, where the
340
- > signed `orgId` and `role` claims are re-derived from the database rather than
341
- > trusted as minted.
337
+ > membership changes, so its stored active organization only counts while a
338
+ > current member row backs it. Remove a member and their org-scoped private-file
339
+ > access stops on the next request — including on the JWT fast path, where the
340
+ > signed `orgId` and `role` claims are re-derived from the database rather than
341
+ > trusted as minted.
342
342
 
343
343
 
344
344
  ## Generated Files
@@ -31,13 +31,13 @@ External Service → POST /webhooks/v1/inbound/stripe
31
31
  ```
32
32
 
33
33
  > **The body cap comes before signature verification, deliberately.** The
34
- > signature *is* the authentication on this route, and it cannot be checked
35
- > until the body has been read — so the read itself carries the limit.
36
- > Without it, an unsigned caller would choose how much the Worker buffers,
37
- > hashes, stores and enqueues before being turned away. Bodies over 1 MiB get
38
- > `413`, distinct from the `401` an invalid signature returns.
34
+ > signature *is* the authentication on this route, and it cannot be checked
35
+ > until the body has been read — so the read itself carries the limit.
36
+ > Without it, an unsigned caller would choose how much the Worker buffers,
37
+ > hashes, stores and enqueues before being turned away. Bodies over 1 MiB get
38
+ > `413`, distinct from the `401` an invalid signature returns.
39
39
  >
40
- > The same cap applies to actions declaring `webhook:` verification.
40
+ > The same cap applies to actions declaring `webhook:` verification.
41
41
 
42
42
 
43
43
  The API returns `200 OK` immediately after enqueuing. Your handler logic runs asynchronously via the queue consumer, so external services don't time out waiting for your processing to complete.
@@ -30,6 +30,14 @@ Provision hands back a one-time approval that finalize must present, so the uplo
30
30
 
31
31
  A typical deploy takes one to two minutes, most of it compilation.
32
32
 
33
+ ## Naming the Worker
34
+
35
+ The card's **Worker name** field seeds itself from your project's `name` and decides what the deploy is called. It appears as soon as your project compiles green — before you connect Cloudflare — so naming the app isn't gated behind an OAuth round trip. That single name drives the Worker **and** every resource derived from it — `<name>-auth`, `<name>-features`, and any audit, webhook, or file-metadata databases your features pull in.
36
+
37
+ Start remembers the name per chat, so redeploying the same conversation updates the same app in place. Editing it deploys a **separate copy**: a new Worker, on new databases, at a new URL. The previous app keeps running untouched — nothing is renamed or deleted in your Cloudflare account, so a name you've finished with is yours to clean up in the dashboard.
38
+
39
+ Names follow Cloudflare's rule — lowercase letters, digits, hyphens, and underscores, starting with a letter or digit, up to 64 characters. The field normalizes as you type and blocks the deploy rather than letting the broker reject the name a minute in.
40
+
33
41
  ## What deploys
34
42
 
35
43
  Supported today, end to end:
@@ -137,12 +145,25 @@ Click Deploy as often as you like:
137
145
  - **Assets are content-addressed** — a redeploy with unchanged CMS/Account assets uploads zero bytes.
138
146
  - **Realtime classes register once** — Durable Object registrations follow Cloudflare's append-only chronology; a redeploy sends nothing new.
139
147
 
148
+ > **Changing your schema between browser deploys.** Start keeps your project sources, not the migration history the compiler generates from them — so editing a table and redeploying regenerates the *initial* migration instead of adding an incremental one. Against a database that already ran the original, the ledger sees the same migration with different content and refuses the deploy:
149
+ >
150
+ > ```
151
+ > migration "quickback/drizzle/features/0000_quickback.sql" already ran against this
152
+ > database with different content. Applying it again would leave the schema in a
153
+ > state neither version describes.
154
+ > ```
155
+ >
156
+ > That refusal is the safe outcome — it stops before touching your schema. To move past it, either give the deploy a new name (a fresh app on fresh databases), or [download the project](/start/deploy) and use `quickback deploy`, which keeps `quickback/drizzle/` alongside your sources and generates proper incremental migrations. Evolving a deployed schema in place is a CLI workflow today.
157
+
158
+
140
159
  ## After it's live
141
160
 
161
+ Start keeps a chip above the composer naming the app this chat deployed, with links straight into its CMS, Account UI, and API. It is read from local storage rather than the deploy card's live state, so it survives a reload — refreshing no longer loses the way back to your app.
162
+
142
163
  Your app answers at `https://<app-name>.<your-subdomain>.workers.dev`:
143
164
 
144
165
  - **API-only apps** serve a small landing page at `/` (browsers) and a JSON service descriptor (everything else), linking `/api/v1`, `/openapi.json`, and `/health` — never a bare 404 on first visit.
145
- - **Apps with the Account UI or CMS** redirect `/` to the login page or CMS.
166
+ - **Apps with the Account UI or CMS** redirect `/` to the login page or CMS. Both ship by default from Start, at `/cms/` and `/account/`. The CMS needs `user.role === "appmanager"`, which your first deploy seeds for you — see [Owner access](/start/deploy-owner-access).
146
167
  - `/health` reports status and whether email is configured.
147
168
 
148
169
  The deployed app is a normal Worker in *your* account: logs, metrics, and settings live in your Cloudflare dashboard, and you can take over management with the CLI at any time — download the project from Start and `quickback deploy` picks up the same resources by name.
@@ -151,6 +172,7 @@ The deployed app is a normal Worker in *your* account: logs, metrics, and settin
151
172
 
152
173
  - **Owner login in deployed apps** ("Continue with Cloudflare" as the app's admin sign-in) — the UI half ships with the Account SPA, but the server half and its per-app OAuth client are still in progress.
153
174
  - **Custom domains** for browser-deployed apps — `workers.dev` only for now.
175
+ - **Schema changes to an already-deployed app** — browser deploys carry no migration history, so an evolving schema needs the CLI (see the callout above).
154
176
  - The blocked resources listed above.
155
177
 
156
178
  These all work — or will work first — through [`quickback deploy`](/start/deploy).
@@ -121,10 +121,10 @@ an asset is served only when the Worker fetches it. That is what keeps
121
121
  `publish: 'admin'` spec is not a public URL. Do not relax it.
122
122
 
123
123
  > **Reading the spec from disk.** Tooling that expected `openapi.json` at the
124
- > project root — `npx openapi-typescript openapi.json`, CI steps, lint rules —
125
- > must point at `src/apps/__quickback/spec/openapi.json`. `api-types.gen.ts` is
126
- > unaffected: the compiler still generates it at the project root for you, so
127
- > most projects never need to run `openapi-typescript` by hand.
124
+ > project root — `npx openapi-typescript openapi.json`, CI steps, lint rules —
125
+ > must point at `src/apps/__quickback/spec/openapi.json`. `api-types.gen.ts` is
126
+ > unaffected: the compiler still generates it at the project root for you, so
127
+ > most projects never need to run `openapi-typescript` by hand.
128
128
 
129
129
 
130
130
  ## Retired Paths
@@ -422,8 +422,8 @@ Only `meta/_journal.json` is expected to change, because that is where a new
422
422
  entry lands.
423
423
 
424
424
  > If `quickback migrations doctor` reports snapshot drift — more journal entries
425
- > than snapshot files — that drift came from your project's history, not from
426
- > compiling. Compile does not delete snapshots it was never given.
425
+ > than snapshot files — that drift came from your project's history, not from
426
+ > compiling. Compile does not delete snapshots it was never given.
427
427
 
428
428
 
429
429
  For non-interactive environments (cloud compile, CI), table/column renames must be declared with `compiler.migrations.renames` in `quickback.config.ts`. This avoids interactive rename prompts during migration generation.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@quickback-dev/cli",
3
- "version": "0.60.6",
3
+ "version": "0.60.7",
4
4
  "description": "CLI for Quickback — define your backend in TypeScript, compile to a fully wired Cloudflare Worker (or Bun / Node) with auth, database, migrations, security, CMS, and Account UI.",
5
5
  "author": "Kardoe Incorporated <paul@kardoe.com>",
6
6
  "license": "MIT",
@@ -108,16 +108,16 @@ under the same `openapi.publish` gate. See [Output
108
108
  structure](/tooling/output#spec-blobs-are-assets-not-bundle).
109
109
 
110
110
  > **BREAKING — `openapi.json` is no longer written to the project root.**
111
- > Tooling that read it (`npx openapi-typescript openapi.json`, CI steps, lint
112
- > rules) must point at `src/apps/__quickback/spec/openapi.json`.
111
+ > Tooling that read it (`npx openapi-typescript openapi.json`, CI steps, lint
112
+ > rules) must point at `src/apps/__quickback/spec/openapi.json`.
113
113
  >
114
- > `api-types.gen.ts` is **unaffected** — the compiler still generates it at the
115
- > project root, which is why most projects never invoke `openapi-typescript`
116
- > directly and need no change at all.
114
+ > `api-types.gen.ts` is **unaffected** — the compiler still generates it at the
115
+ > project root, which is why most projects never invoke `openapi-typescript`
116
+ > directly and need no change at all.
117
117
  >
118
- > On your first compile after upgrading, the CLI deletes the now-orphaned root
119
- > copies. It only does so for blobs it just re-emitted under the new path, so an
120
- > older compiler that still writes to the root is never second-guessed.
118
+ > On your first compile after upgrading, the CLI deletes the now-orphaned root
119
+ > copies. It only does so for blobs it just re-emitted under the new path, so an
120
+ > older compiler that still writes to the root is never second-guessed.
121
121
 
122
122
 
123
123
  ## v0.59.1 — July 29, 2026
@@ -131,34 +131,34 @@ Null-safe scope semantics make this safe (these apply **inside `all`/`any` group
131
131
  - **Owner** (`ctx.userId`) and **soft-delete** (`isNull deletedAt`) are always present / always AND'd — never relaxed.
132
132
 
133
133
  > **`ctx.activeTeamId` is implicitly optional only while another tenant arm
134
- > bounds the query.** Dropping the team filter is safe when an org (or owner,
135
- > `via:`, `permission`, or a binding group) arm still constrains the rows —
136
- > that is what makes it "the teamless rest-of-org slice".
134
+ > bounds the query.** Dropping the team filter is safe when an org (or owner,
135
+ > `via:`, `permission`, or a binding group) arm still constrains the rows —
136
+ > that is what makes it "the teamless rest-of-org slice".
137
137
  >
138
- > When **team is the only tenant arm**, there is no rest-of-org slice to fall
139
- > back to, so an absent claim denies with ``sql`1 = 0` `` instead of dropping.
140
- > Otherwise every predicate would evaluate to `undefined`, the firewall would
141
- > return no `WHERE` at all, and any authenticated caller would read every
142
- > tenant's rows.
138
+ > When **team is the only tenant arm**, there is no rest-of-org slice to fall
139
+ > back to, so an absent claim denies with ``sql`1 = 0` `` instead of dropping.
140
+ > Otherwise every predicate would evaluate to `undefined`, the firewall would
141
+ > return no `WHERE` at all, and any authenticated caller would read every
142
+ > tenant's rows.
143
143
  >
144
- > This matters because `ctx.activeTeamId` is absent in ordinary
145
- > configurations: the Better Auth teams plugin is opt-in, and the external auth
146
- > provider does not supply the claim at all. Note that soft-delete and literal
147
- > predicates (`kind = 'public'`) are **row filters, not tenant boundaries** —
148
- > they do not count as a co-existing arm.
144
+ > This matters because `ctx.activeTeamId` is absent in ordinary
145
+ > configurations: the Better Auth teams plugin is opt-in, and the external auth
146
+ > provider does not supply the claim at all. Note that soft-delete and literal
147
+ > predicates (`kind = 'public'`) are **row filters, not tenant boundaries** —
148
+ > they do not count as a co-existing arm.
149
149
  >
150
- > A table with a `teamId` column and no `organizationId` gets team-only
151
- > isolation from auto-detection, without any explicit `firewall` block.
150
+ > A table with a `teamId` column and no `organizationId` gets team-only
151
+ > isolation from auto-detection, without any explicit `firewall` block.
152
152
  >
153
- > **You will usually see this as a compile error rather than a silent deny.**
154
- > When the teams sub-plugin is off, `ctx.activeTeamId` can never be populated,
155
- > so a team-only firewall would deny *every* request — the compiler rejects
156
- > that configuration up front and tells you to either enable teams
157
- > (`auth: { plugins: { organization: { teams: { enabled: true } } } }`, which
158
- > requires object-form plugin config) or give the resource a bindable tenant
159
- > scope. The runtime deny remains the backstop for the case the compiler can't
160
- > see, such as teams being enabled while a particular caller has not selected
161
- > a team.
153
+ > **You will usually see this as a compile error rather than a silent deny.**
154
+ > When the teams sub-plugin is off, `ctx.activeTeamId` can never be populated,
155
+ > so a team-only firewall would deny *every* request — the compiler rejects
156
+ > that configuration up front and tells you to either enable teams
157
+ > (`auth: { plugins: { organization: { teams: { enabled: true } } } }`, which
158
+ > requires object-form plugin config) or give the resource a bindable tenant
159
+ > scope. The runtime deny remains the backstop for the case the compiler can't
160
+ > see, such as teams being enabled while a particular caller has not selected
161
+ > a team.
162
162
 
163
163
 
164
164
  `nullable: true` is a separate database-schema opt-in, supported only on an
@@ -222,17 +222,17 @@ changeset writes; without it the compiler warns that changeset writes to a
222
222
  A table that declares no `read.access` is authenticated-only, identical to
223
223
  its CRUD read.
224
224
 
225
- > **This differs from `?include=` on purpose.** FK-graph `?include=`
226
- > *requires* its target to declare `read.access` and is a compile error
227
- > otherwise; a view surface does not.
225
+ > **This differs from `?include=` on purpose.** FK-graph `?include=`
226
+ > *requires* its target to declare `read.access` and is a compile error
227
+ > otherwise; a view surface does not.
228
228
  >
229
- > Neither surface is weaker than the other. A table with no declared
230
- > `read.access` is already authenticated-only-plus-firewall when read
231
- > directly through CRUD, so embedding it in a view grants nothing extra — the
232
- > view is exactly as strong as reading the table itself, which is the rule
233
- > both surfaces are held to. The difference is authoring strictness, not
234
- > enforcement: `?include=` makes you state the target's read policy before
235
- > embedding it, while a view inherits whatever the table already does.
229
+ > Neither surface is weaker than the other. A table with no declared
230
+ > `read.access` is already authenticated-only-plus-firewall when read
231
+ > directly through CRUD, so embedding it in a view grants nothing extra — the
232
+ > view is exactly as strong as reading the table itself, which is the rule
233
+ > both surfaces are held to. The difference is authoring strictness, not
234
+ > enforcement: `?include=` makes you state the target's read policy before
235
+ > embedding it, while a view inherits whatever the table already does.
236
236
 
237
237
 
238
238
  {/* doc-compile: skip — illustrative fragment: the point is the `read.access` line and the three 403s below it, so the host `customers` table and its columns are deliberately not repeated. Whole-file table examples live in /define/schema. */}
@@ -146,35 +146,35 @@ POST /storage/v1/bucket
146
146
  | `user` | Owner only | Owner only |
147
147
 
148
148
  > **`writeScope: "user"` is enforced per object, not per organization.** Any
149
- > member of the org may upload to such a bucket, so the object key is the only
150
- > thing separating one member's files from another's — and keys are built from a
151
- > client-supplied path. Before any write, Quickback checks whether the target key
152
- > already has an owner: writing to a key owned by another user returns `403`
153
- > `ACCESS_OWNERSHIP_REQUIRED`.
149
+ > member of the org may upload to such a bucket, so the object key is the only
150
+ > thing separating one member's files from another's — and keys are built from a
151
+ > client-supplied path. Before any write, Quickback checks whether the target key
152
+ > already has an owner: writing to a key owned by another user returns `403`
153
+ > `ACCESS_OWNERSHIP_REQUIRED`.
154
154
  >
155
- > This applies to presigned uploads too. A signed `PUT` URL *is* a write
156
- > capability, so ownership is checked **before** the URL is issued, not when the
157
- > bytes land.
155
+ > This applies to presigned uploads too. A signed `PUT` URL *is* a write
156
+ > capability, so ownership is checked **before** the URL is issued, not when the
157
+ > bytes land.
158
158
 
159
159
 
160
160
  > **The read policy covers metadata, not just bytes.** `readScope` and
161
- > `readRoles` are applied on every route that can reveal an object — downloads,
162
- > `GET /object` listings, and `POST /url` lookups alike. Being a member of the
163
- > owning organization is not sufficient on its own.
161
+ > `readRoles` are applied on every route that can reveal an object — downloads,
162
+ > `GET /object` listings, and `POST /url` lookups alike. Being a member of the
163
+ > owning organization is not sufficient on its own.
164
164
  >
165
- > - **Listing** (`GET /object`) returns only objects in buckets you may read.
166
- > The filter is applied in the query, so `limit` and `offset` page over your
167
- > readable set — you never receive a short page because rows were removed
168
- > after the fact. Naming a bucket you cannot read returns `403`
169
- > `ACCESS_ROLE_REQUIRED`; with no `bucket` filter, unreadable buckets are
170
- > simply absent.
171
- > - **URL lookup** (`POST /url`) returns `404` when the bucket policy denies
172
- > you — the same response as an object that does not exist. This is
173
- > deliberate: a `403` would confirm that the id or key you supplied resolves
174
- > to a real object, letting a caller enumerate objects they cannot read.
165
+ > - **Listing** (`GET /object`) returns only objects in buckets you may read.
166
+ > The filter is applied in the query, so `limit` and `offset` page over your
167
+ > readable set — you never receive a short page because rows were removed
168
+ > after the fact. Naming a bucket you cannot read returns `403`
169
+ > `ACCESS_ROLE_REQUIRED`; with no `bucket` filter, unreadable buckets are
170
+ > simply absent.
171
+ > - **URL lookup** (`POST /url`) returns `404` when the bucket policy denies
172
+ > you — the same response as an object that does not exist. This is
173
+ > deliberate: a `403` would confirm that the id or key you supplied resolves
174
+ > to a real object, letting a caller enumerate objects they cannot read.
175
175
  >
176
- > For `readScope: "user"` buckets, both routes additionally require that you own
177
- > the object.
176
+ > For `readScope: "user"` buckets, both routes additionally require that you own
177
+ > the object.
178
178
 
179
179
 
180
180
  ### Role-Based Access
@@ -334,11 +334,11 @@ The files worker:
334
334
  4. Serves the file or returns 403
335
335
 
336
336
  > Step 2 is a live lookup, not a read of the session row. A session outlives
337
- > membership changes, so its stored active organization only counts while a
338
- > current member row backs it. Remove a member and their org-scoped private-file
339
- > access stops on the next request — including on the JWT fast path, where the
340
- > signed `orgId` and `role` claims are re-derived from the database rather than
341
- > trusted as minted.
337
+ > membership changes, so its stored active organization only counts while a
338
+ > current member row backs it. Remove a member and their org-scoped private-file
339
+ > access stops on the next request — including on the JWT fast path, where the
340
+ > signed `orgId` and `role` claims are re-derived from the database rather than
341
+ > trusted as minted.
342
342
 
343
343
 
344
344
  ## Generated Files
@@ -31,13 +31,13 @@ External Service → POST /webhooks/v1/inbound/stripe
31
31
  ```
32
32
 
33
33
  > **The body cap comes before signature verification, deliberately.** The
34
- > signature *is* the authentication on this route, and it cannot be checked
35
- > until the body has been read — so the read itself carries the limit.
36
- > Without it, an unsigned caller would choose how much the Worker buffers,
37
- > hashes, stores and enqueues before being turned away. Bodies over 1 MiB get
38
- > `413`, distinct from the `401` an invalid signature returns.
34
+ > signature *is* the authentication on this route, and it cannot be checked
35
+ > until the body has been read — so the read itself carries the limit.
36
+ > Without it, an unsigned caller would choose how much the Worker buffers,
37
+ > hashes, stores and enqueues before being turned away. Bodies over 1 MiB get
38
+ > `413`, distinct from the `401` an invalid signature returns.
39
39
  >
40
- > The same cap applies to actions declaring `webhook:` verification.
40
+ > The same cap applies to actions declaring `webhook:` verification.
41
41
 
42
42
 
43
43
  The API returns `200 OK` immediately after enqueuing. Your handler logic runs asynchronously via the queue consumer, so external services don't time out waiting for your processing to complete.
@@ -30,6 +30,14 @@ Provision hands back a one-time approval that finalize must present, so the uplo
30
30
 
31
31
  A typical deploy takes one to two minutes, most of it compilation.
32
32
 
33
+ ## Naming the Worker
34
+
35
+ The card's **Worker name** field seeds itself from your project's `name` and decides what the deploy is called. It appears as soon as your project compiles green — before you connect Cloudflare — so naming the app isn't gated behind an OAuth round trip. That single name drives the Worker **and** every resource derived from it — `<name>-auth`, `<name>-features`, and any audit, webhook, or file-metadata databases your features pull in.
36
+
37
+ Start remembers the name per chat, so redeploying the same conversation updates the same app in place. Editing it deploys a **separate copy**: a new Worker, on new databases, at a new URL. The previous app keeps running untouched — nothing is renamed or deleted in your Cloudflare account, so a name you've finished with is yours to clean up in the dashboard.
38
+
39
+ Names follow Cloudflare's rule — lowercase letters, digits, hyphens, and underscores, starting with a letter or digit, up to 64 characters. The field normalizes as you type and blocks the deploy rather than letting the broker reject the name a minute in.
40
+
33
41
  ## What deploys
34
42
 
35
43
  Supported today, end to end:
@@ -137,12 +145,25 @@ Click Deploy as often as you like:
137
145
  - **Assets are content-addressed** — a redeploy with unchanged CMS/Account assets uploads zero bytes.
138
146
  - **Realtime classes register once** — Durable Object registrations follow Cloudflare's append-only chronology; a redeploy sends nothing new.
139
147
 
148
+ > **Changing your schema between browser deploys.** Start keeps your project sources, not the migration history the compiler generates from them — so editing a table and redeploying regenerates the *initial* migration instead of adding an incremental one. Against a database that already ran the original, the ledger sees the same migration with different content and refuses the deploy:
149
+ >
150
+ > ```
151
+ > migration "quickback/drizzle/features/0000_quickback.sql" already ran against this
152
+ > database with different content. Applying it again would leave the schema in a
153
+ > state neither version describes.
154
+ > ```
155
+ >
156
+ > That refusal is the safe outcome — it stops before touching your schema. To move past it, either give the deploy a new name (a fresh app on fresh databases), or [download the project](/start/deploy) and use `quickback deploy`, which keeps `quickback/drizzle/` alongside your sources and generates proper incremental migrations. Evolving a deployed schema in place is a CLI workflow today.
157
+
158
+
140
159
  ## After it's live
141
160
 
161
+ Start keeps a chip above the composer naming the app this chat deployed, with links straight into its CMS, Account UI, and API. It is read from local storage rather than the deploy card's live state, so it survives a reload — refreshing no longer loses the way back to your app.
162
+
142
163
  Your app answers at `https://<app-name>.<your-subdomain>.workers.dev`:
143
164
 
144
165
  - **API-only apps** serve a small landing page at `/` (browsers) and a JSON service descriptor (everything else), linking `/api/v1`, `/openapi.json`, and `/health` — never a bare 404 on first visit.
145
- - **Apps with the Account UI or CMS** redirect `/` to the login page or CMS.
166
+ - **Apps with the Account UI or CMS** redirect `/` to the login page or CMS. Both ship by default from Start, at `/cms/` and `/account/`. The CMS needs `user.role === "appmanager"`, which your first deploy seeds for you — see [Owner access](/start/deploy-owner-access).
146
167
  - `/health` reports status and whether email is configured.
147
168
 
148
169
  The deployed app is a normal Worker in *your* account: logs, metrics, and settings live in your Cloudflare dashboard, and you can take over management with the CLI at any time — download the project from Start and `quickback deploy` picks up the same resources by name.
@@ -151,6 +172,7 @@ The deployed app is a normal Worker in *your* account: logs, metrics, and settin
151
172
 
152
173
  - **Owner login in deployed apps** ("Continue with Cloudflare" as the app's admin sign-in) — the UI half ships with the Account SPA, but the server half and its per-app OAuth client are still in progress.
153
174
  - **Custom domains** for browser-deployed apps — `workers.dev` only for now.
175
+ - **Schema changes to an already-deployed app** — browser deploys carry no migration history, so an evolving schema needs the CLI (see the callout above).
154
176
  - The blocked resources listed above.
155
177
 
156
178
  These all work — or will work first — through [`quickback deploy`](/start/deploy).
@@ -121,10 +121,10 @@ an asset is served only when the Worker fetches it. That is what keeps
121
121
  `publish: 'admin'` spec is not a public URL. Do not relax it.
122
122
 
123
123
  > **Reading the spec from disk.** Tooling that expected `openapi.json` at the
124
- > project root — `npx openapi-typescript openapi.json`, CI steps, lint rules —
125
- > must point at `src/apps/__quickback/spec/openapi.json`. `api-types.gen.ts` is
126
- > unaffected: the compiler still generates it at the project root for you, so
127
- > most projects never need to run `openapi-typescript` by hand.
124
+ > project root — `npx openapi-typescript openapi.json`, CI steps, lint rules —
125
+ > must point at `src/apps/__quickback/spec/openapi.json`. `api-types.gen.ts` is
126
+ > unaffected: the compiler still generates it at the project root for you, so
127
+ > most projects never need to run `openapi-typescript` by hand.
128
128
 
129
129
 
130
130
  ## Retired Paths
@@ -422,8 +422,8 @@ Only `meta/_journal.json` is expected to change, because that is where a new
422
422
  entry lands.
423
423
 
424
424
  > If `quickback migrations doctor` reports snapshot drift — more journal entries
425
- > than snapshot files — that drift came from your project's history, not from
426
- > compiling. Compile does not delete snapshots it was never given.
425
+ > than snapshot files — that drift came from your project's history, not from
426
+ > compiling. Compile does not delete snapshots it was never given.
427
427
 
428
428
 
429
429
  For non-interactive environments (cloud compile, CI), table/column renames must be declared with `compiler.migrations.renames` in `quickback.config.ts`. This avoids interactive rename prompts during migration generation.