## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
Refactor / dependency cleanup.
## What is the current behavior?
`apps/studio` lists both `@supabase/pg-meta` (workspace package) as a
runtime dep and `@supabase/postgres-meta` (external npm package,
`^0.64.4`) as a devDependency. The external package is used only for
type imports across 44 files — there is no runtime usage and no codegen
pipeline that needs it.
## What is the new behavior?
Every `Postgres*` type import (`PostgresTable`, `PostgresColumn`,
`PostgresPolicy`, `PostgresTrigger`, `PostgresView`,
`PostgresMaterializedView`, `PostgresForeignTable`, `PostgresSchema`,
`PostgresPublication`, `PostgresRelationship`, `PostgresPrimaryKey`) is
replaced with its `PG*` counterpart from `@supabase/pg-meta`, and the
external dep is removed from \`apps/studio/package.json\`. Top-level
type re-exports were added to \`packages/pg-meta/src/index.ts\` so
consumers can import directly from the package root.
Two latent issues surfaced by the stricter pg-meta types are also fixed:
- \`data/foreign-tables/foreign-tables-query.ts\` was casting
foreign-table results as \`PostgresView[]\`; corrected to
\`PGForeignTable[]\`.
- \`pg-meta\`'s \`PGTrigger\` Zod schema declared
\`orientation\`/\`activation\` as \`z.string()\`, inconsistent with
pg-meta's own \`getDatabaseTriggerUpdateSQL\` helper that requires the
narrow literal unions; tightened to \`z.enum\`.
## Additional context
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Chores**
* Updated internal TypeScript type definitions across the codebase to
use the latest type system from `@supabase/pg-meta`.
* Removed `@supabase/postgres-meta` dependency.
* Enhanced type validation for database triggers and schemas to enforce
stricter constraints.
[](https://app.coderabbit.ai/change-stack/supabase/supabase/pull/45844)
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Mark provenance of SQL via the branded types SafeSqlFragment and
UntrustedSqlFragment. Only SafeSqlFragment should be executed;
UntrustedSqlFragments require some kind of implicit user approval (show
on screen + user has to click something) before they are promoted to
SafeSqlFragment.
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **New Features**
* Editor and RLS tester show loading states for inferred/generated SQL
and include a dedicated user SQL editor for safer edits.
* **Refactor**
* Platform-wide SQL handling tightened: snippets and AI-generated SQL
are treated as untrusted/display-only until promoted, improving safety
and consistency.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Fixes the RLS policies page showing self-contradictory or wrong
admonitions for tables with partial grants. Classifies each table using
the same `granted / custom / revoked` semantics used by the Data API
settings page so the two views agree on what counts as "exposed".
**Changed:**
- `PolicyTableRow` now uses `useTableApiAccessQuery` (shared cache with
the Table Editor sidebar) instead of a bespoke
`tables-roles-access-query`
- Boolean soup collapsed into a single `TableDataApiStatus`
discriminated union (`schema-not-exposed | no-grants | custom-grants |
publicly-readable | locked-by-rls | secured`) via a pure helper
- Admonition copy for `no-grants` and `locked-by-rls` updated; a table
with no policies but full grants now reads "No data will be returned via
the Data API as no RLS policies exist on this table." instead of the
earlier self-contradictory "can be accessed but no data will be
returned"
- `table-api-access-query.ts` now exposes a `grantStatus: 'granted' |
'custom'` on `access` entries — `granted` = all 3 API roles × all 4 CRUD
privileges (matches `getTableGrantsCTEs` in pg-meta)
**Added:**
- New `custom-grants` admonition: "This table has custom Data API
permissions — access may be restricted for some roles or operations."
- Unit tests for `getTableDataApiStatus`, `getTableAdmonitionMessage`,
and `isFullyGranted`
**Removed:**
- `data/tables/tables-roles-access-query.ts` and the `rolesAccess` key —
no more callers
## To test
On a project with the `public` schema exposed, for each scenario check
the admonition shown on `/project/{ref}/auth/policies`:
1. Table with full standard grants, RLS on, no policies → "No data will
be returned via the Data API as no RLS policies exist on this table."
2. Table with full standard grants, RLS off → yellow warning "can be
accessed by anyone"
3. Table with partial grants (e.g. only `GRANT SELECT ON t TO anon`) →
new "custom Data API permissions" admonition regardless of RLS state
4. Table with no anon/authenticated/service_role grants → "cannot be
accessed via the Data API"
5. Schema not in the exposed list → "schema not exposed" admonition with
link
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Tests**
* Added unit tests covering table Data API/RLS status classification and
API grant validation.
* **Refactor**
* Introduced a unified per-table API/RLS status model and reusable
utilities to derive display status and admonitions.
* Simplified UI logic to drive access indicators and warnings from the
new status.
* **Chores**
* Removed legacy role-based access query and its related keying logic.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
---------
Co-authored-by: Alaister Young <10985857+alaister@users.noreply.github.com>
* start
* added panels
* remove stuff
* fixes and refinements
* clean up
* remove old assistant panel
* resizable assistant kinda
* use icon
* Add missing package
* remove canvas
* add suggestions
* updated empty state if no tables exist
* fix table condition
* Implement diffing if using assistant in sql editor
* Reinstate old assistant in SQL editor if feature preview is off
* pane size adjustment
* assistant button corners
* Add SQL snippet content to assistant if opening assistant in sql editor
* Add the necessary checks for opt in and hipaa
* revert adding snippet to assistant when opening assistant in sql editor
* Add cmd i shortcut
* Add admonitions for when disablePrompt is toggled on, and if no api key is set. Add footer note RE rate limitation
* Bump ai package in packages
* some fixes for backwards compability depending on feature preview toggled
* Rename feature preview property for new assistant
* Smol fix
* Prevent SQL snippet from running until message is finished
* only loading last message
* fix z-index
* save chat state to global state
* add debug to failed ai queries
* Add basic contextual invalidation
* Add explain code action to SQL editor
* Add link to abort ongoing queries from SqlSnippet
* Update feature preview content
* Fix
* Fix
* Fix
* Te4st
* Fix tests
* ONly show ai button within a project
* Fix PH tracking
* Beef up a bit more event tracking
* Rough fix to padding when assistant is open
* A bit more telemetry stuff
* Update prompts
* fix rls editing via assistant
* Update generate-v3.ts
prompt to get auth schema too
* Add policy satement to assistant when editing
* Address all comments
* fixc
* Fix SqlSnippet not taking full width on larger viewports
* Adjust max width
---------
Co-authored-by: Saxon Fletcher <saxonafletcher@gmail.com>
