## Context
This PR adds the core structure for RLS implementation:
- RLS data model
- RLS service layer
- RLS WorkspaceMigration and Syncable Entity + cache + Validations
- RLS resolver layer
- ORM layer with RLS Predicate to ORM WHERE clause conversion with
workspaceMember record transposition
Tests are missing though
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> Establishes core row-level permissions infrastructure and enforcement
across the stack.
>
> - Backend: new `rowLevelPermissionPredicate` and
`rowLevelPermissionPredicateGroup` entities, TypeORM migration, feature
flag `IS_ROW_LEVEL_PERMISSION_PREDICATES_ENABLED`, flat-entity
maps/cache wiring, services and GraphQL resolvers for CRUD, and
inclusion of `workspaceMember` in auth context
> - ORM: applies row-level permission predicates to SELECT, DELETE, and
SOFT DELETE query builders; propagates context through
GlobalWorkspaceOrmManager/EntityManager
> - GraphQL: generated schema/types/queries/mutations for
creating/updating/deleting/fetching predicates and groups
> - Frontend: settings page adds a gated "Record-level" section
(placeholder) and metadata error handler labels for new entities
>
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
fe955cc4588a92157afa6795fb574189a4be1e93. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
## Summary
This PR migrates workflow CRUD operations to properly use the Common API
layer's authentication context, addressing the issues from the reverted
PR #15875.
The original PR was reverted because the Common API required passing
either a User or an API Key for authentication, which was problematic
for workflows. Since then, the "Application" concept was introduced in
the Common API layer, allowing for token injection in serverless
functions.
This PR leverages the "Twenty Standard Application" concept for
non-manual workflow triggers, providing a clean authentication path
without the issues of user impersonation.
## Changes
### Core Infrastructure
- **RecordCrudExecutionContext**: Replace `workspaceId` with full
`authContext`
- **WorkflowExecutionContext**: Add `authContext` field to carry
authentication info
- **ToolGeneratorContext/ToolSpecification**: Add optional `authContext`
support for tool generation
### Authentication Flow
- **WorkflowExecutionContextService**: Build appropriate auth context
based on trigger type:
- **Manual triggers**: Use user's workspace auth context with their role
permissions
- **Non-manual triggers**: Use Twenty Standard Application auth context
(bypasses permission checks or uses default serverless function role)
- **ApplicationService**: Add `findTwentyStandardApplicationOrThrow`
method to retrieve the system application
- **UserWorkspaceService**: Make relations configurable in
`getUserWorkspaceForUserOrThrow` to load only what's needed
### CRUD Services Migration
All 5 record CRUD services now receive `authContext` instead of
`workspaceId`:
- `CreateRecordService`
- `UpdateRecordService`
- `DeleteRecordService`
- `FindRecordsService`
- `UpsertRecordService`
### Workflow Actions
All record CRUD workflow actions pass `executionContext.authContext` to
the services:
- `CreateRecordWorkflowAction`
- `UpdateRecordWorkflowAction`
- `DeleteRecordWorkflowAction`
- `FindRecordsWorkflowAction`
- `UpsertRecordWorkflowAction`
### AI Agent Integration
- AI agent workflow action passes auth context to agent executor
- Tool provider and MCP protocol service support auth context
propagation
## Benefits
- ✅ Proper authentication for workflow CRUD operations via Common API
- ✅ Non-manual triggers use system application context (no user
impersonation issues)
- ✅ Manual triggers preserve user permissions correctly
- ✅ Foundation for better permission handling in automated workflows
- ✅ Cleaner separation between user-initiated and system-initiated
operations
## Related
- Reverted PR: #15875
## Context
This PR refactors how authentication context is handled in the GraphQL
resolver layer. Previously, the auth context was baked into the schema
builder context at schema creation time. Now, the auth context is
extracted from the request at query execution time, enabling better
schema caching.
## Implementation
- Removed user-specific data from schema cache key: The schema cache key
no longer includes userId and apiKeyId, allowing the same schema to be
shared across all users in a workspace
- Cache key simplified from
${workspaceId}-${userId}-${apiKeyId}-${url}-${cacheVersion} to
${workspaceId}-${url}-${cacheVersion}
- Removed authContext from WorkspaceSchemaBuilderContext: The schema
builder context is now purely about object/field metadata, not about who
is accessing it
- Created createQueryRunnerContext utility: A new helper that combines
the schema builder context with the auth context from the actual request
- Updated all resolver factories: All 15 resolver factories now use
createQueryRunnerContext to build the full context needed by query
runners at execution time
## Benefits
- Improved cache efficiency: GraphQL schemas are now cached per
workspace rather than per user, significantly reducing memory usage and
schema regeneration
- Cleaner separation of concerns: Schema building (metadata-driven) is
now separate from query execution (auth-driven)
- Fresh auth context: Auth context is always retrieved from the current
request, ensuring it reflects the latest state
- Reduced complexity: Simplified the schema creation path by removing
unnecessary auth checks
## Next steps
- Introduce a hash in redis and expose it in the req object so the yoga
patch can access the hash and use it during its own cache key
computation. This way we will be able to invalidate the local cache in
yoga graphql schema generation without restarting pods.
Two challenges with error messages
- always provide a useful/meaningful error message for the end user
instead of the generic one. eg: show "Wrong password" and not "An error
occured"
- avoid technical details unless error regards a technical feature. eg:
show "An error occured" and not "Invalid post-hook payload."; but do
show "Invalid issuer URL." as it occurs while configuring SSO
What this PR does
- Make userFriendlyMessage mandatory for widely used
GraphqlQueryRunnerException and CommonQueryRunnerException, so that
developers are forced to ask themselves what the error message should
be, and as it contains very wide error codes (eg: "Bad request") which
should not be mapped to just one default message
- Keep userFriendlyMessage optional for service-specific exceptions (eg:
workflowStepExecutorException), but convert the error code to
userFriendlyMessage mapper to a switch case function with a typecheck
ensuring that all codes are mapped to a message. These default messages
are still overridable where they are thrown.
<!-- CURSOR_SUMMARY -->
> [!NOTE]
> Implements last-modifier tracking and unified actor injection.
>
> - New `ActorFromAuthContextService` replaces createdBy-only logic;
pre-query hooks now inject both `createdBy` and `updatedBy` on create
and `updatedBy` on update
> - Add `updatedBy` standard field to core/custom objects (e.g.,
`person`, `company`, `task`, `note`, `attachment`, `dashboard`,
`workflow`, `workflowRun`) with IDs, metadata builders, and ORM entity
fields
> - Record CRUD: `create` now sets both `createdBy` and `updatedBy`;
`update` sets `updatedBy`; workflow actions use a shared workflow actor
builder; REST base handler uses the new actor service
> - Seeder data and snapshots updated to include `updatedBy`; GraphQL
result formatting adds defaults/handling for empty composite fields
(incl. actor)
> - Frontend `useUpdateOneRecord` upserts returned record into the local
store after mutation
>
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
1f283778e9cb28a35bf84632a1a2997250bb3131. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
Co-authored-by: Félix Malfait <felix@twenty.com>
Fixes https://github.com/twentyhq/twenty/issues/16110
This PR implements Temporal to replace the legacy Date object, in all
features that are time zone sensitive. (around 80% of the app)
Here we define a few utils to handle Temporal primitives and obtain an
easier DX for timezone manipulation, front end and back end.
This PR deactivates the usage of timezone from the graph configuration,
because for now it's always UTC and is not really relevant, let's handle
that later.
Workflows code and backend only code that don't take user input are
using UTC time zone, the affected utils have not been refactored yet
because this PR is big enough.
# New way of filtering on date intervals
As we'll progressively rollup Temporal everywhere in the codebase and
remove `Date` JS object everywhere possible, we'll use the way to filter
that is recommended by Temporal.
This way of filtering on date intervals involves half-open intervals,
and is the preferred way to avoid edge-cases with DST and smallest time
increment edge-case.
## Filtering endOfX with DST edge-cases
Some day-light save time shifts involve having no existing hour, or even
day on certain days, for example Samoa Islands have no 30th of December
2011 : https://www.timeanddate.com/news/time/samoa-dateline.html, it
jumps from 29th to 31st, so filtering on `< next period start` makes it
easier to let the date library handle the strict inferior comparison,
than filtering on `≤ end of period` and trying to compute manually the
end of the period.
For example for Samoa Islands, is end of day `2011-12-29T23:59:59.999`
or is it `2011-12-30T23:59:59.999` ? If you say I don't need to know and
compute it, because I want everything strictly before
`2011-12-29T00:00:00 + start of next day (according to the library which
knows those edge-cases)`, then you have a 100% deterministic way of
computing date intervals in any timezone, for any day of any year.
Of course the Samoa example is an extreme one, but more common ones
involve DST shifts of 1 hour, which are still problematic on certain
days of the year.
## Computing the exact _end of period_
Having an open interval filtering, with `[included - included]` instead
of half-open `[included - excluded)`, forces to compute the open end of
an interval, which often involves taking an arbitrary unit like minute,
second, microsecond or nanosecond, which will lead to edge-case of
unhandled values.
For example, let's say my code computes endOfDay by setting the time to
`23:59:59.999`, if another library, API, or anything else, ends up
giving me a date-time with another time precision `23:59:59.999999999`
(down to the nanosecond), then this date-time will be filtered out,
while it should not.
The good deterministic way to avoid 100% of those complex bugs is to
create a half-open filter :
`≥ start of period` to `< start of next period`
For example :
`≥ 2025-01-01T00:00:00` to `< 2025-01-02T00:00:00` instead of `≥
2025-01-01T00:00:00` to `≤ 2025-01-01T23:59:59.999`
Because, `2025-01-01T00:00:00` = `2025-01-01T00:00:00.000` =
`2025-01-01T00:00:00.000000` = `2025-01-01T00:00:00.000000000` => no
risk of error in computing start of period
But `2025-01-01T23:59:59` ≠ `2025-01-01T23:59:59.999` ≠
`2025-01-01T23:59:59.999999` ≠ `2025-01-01T23:59:59.999999999` =>
existing risk of error in computing end of period
This is why an half-open interval has no risk of error in computing a
date-time interval filter.
Here is a link to this debate :
https://github.com/tc39/proposal-temporal/issues/2568
> For this reason, we recommend not calculating the exact nanosecond at
the end of the day if it's not absolutely necessary. For example, if
it's needed for <= comparisons, we recommend just changing the
comparison code. So instead of <= zdtEndOfDay your code could be <
zdtStartOfNextDay which is easier to calculate and not subject to the
issue of not knowing which unit is the right one.
>
> [Justin Grant](https://github.com/justingrant), top contributor of
Temporal
## Application to our codebase
Applying this half-open filtering paradigm to our codebase means we
would have to rename `IS_AFTER` to `IS_AFTER_OR_EQUAL` and to keep
`IS_BEFORE` (or even `IS_STRICTLY_BEFORE`) to make this half-open
interval self-explanatory everywhere in the codebase, this will avoid
any confusion.
See the relevant issue :
https://github.com/twentyhq/core-team-issues/issues/2010
In the mean time, we'll keep this operand and add this semantic in the
naming everywhere possible.
## Example with a different user timezone
Example on a graph grouped by week in timezone Pacific/Samoa, on a
computer running on Europe/Paris :
<img width="342" height="511" alt="image"
src="https://github.com/user-attachments/assets/9e7d5121-ecc4-4233-835b-f59293fbd8c8"
/>
Then the associated data in the table view, with our **half-open
date-time filter** :
<img width="804" height="262" alt="image"
src="https://github.com/user-attachments/assets/28efe1d7-d2fc-4aec-b521-bada7f980447"
/>
And the associated SQL query result to see how DATE_TRUNC in Postgres
applies its internal start of week logic :
<img width="709" height="220" alt="image"
src="https://github.com/user-attachments/assets/4d0542e1-eaae-4b4b-afa9-5005f48ffdca"
/>
The associated SQL query without parameters to test in your SQL client :
```SQL
SELECT "opportunity"."closeDate" as "close_date", TO_CHAR(DATE_TRUNC('week', "opportunity"."closeDate", 'Pacific/Samoa') AT TIME ZONE 'Pacific/Samoa', 'YYYY-MM-DD') AS "DATE_TRUNC by week start in timezone Pacific/Samoa", "opportunity"."name" FROM "workspace_1wgvd1injqtife6y4rvfbu3h5"."opportunity" "opportunity" ORDER BY "opportunity"."closeDate" ASC NULLS LAST
```
# Date picker simplification (not in this PR)
Our DatePicker component, which is wrapping `react-datepicker` library
component, is now exposing plain dates as string instead of Date object.
The Date object is still used internally to manage the library
component, but since the date picker calendar is only manipulating plain
dates, there is no need to add timezone management to it, and no need to
expose a handleChange with Date object.
The timezone management relies on date time inputs now.
The modification has been made in a previous PR :
https://github.com/twentyhq/twenty/issues/15377 but it's good to
reference it here.
# Calendar feature refactor
Calendar feature has been refactored to rely on Temporal.PlainDate as
much as possible, while leaving some date-fns utils to avoid re-coding
them.
Since the trick is to use utils to convert back and from Date object in
exec env reliably, we can do it everywhere we need to interface legacy
Date object utils and Temporal related code.
## TimeZone is now shown on Calendar :
<img width="894" height="958" alt="image"
src="https://github.com/user-attachments/assets/231f8107-fad6-4786-b532-456692c20f1d"
/>
## Month picker has been refactored
<img width="503" height="266" alt="image"
src="https://github.com/user-attachments/assets/cb90bc34-6c4d-436d-93bc-4b6fb00de7f5"
/>
Since the days weren't useful, the picker has been refactored to remove
the days.
# Miscellaneous
- Fixed a bug with drag and drop edge-case with 2 items in a list.
# Improvements
## Lots of chained operations
It would be nice to create small utils to avoid repeated chained
operations, but that is how Temporal is designed, a very small set of
primitive operations that allow to compose everything needed. Maybe
we'll have wrappers on top of Temporal in the coming years.
## Creation of Temporal objects is throwing errors
If the input is badly formatted Temporal will throw, we might want to
adopt a global strategy to avoid that.
Example :
```ts
const newPlainDate = Temporal.PlainDate.from('bad-string'); // Will throw
```
Removing the feature flag to enable read on replica for all workspaces.
It will still be possible to toggle off the feature by removing the env
variable `PG_DATABASE_REPLICA_URL`.
## Summary
This PR enforces that all custom exceptions must provide a
`userFriendlyMessage`, ensuring end users always see readable error
messages.
## Changes
### Core Changes
- **`CustomException` simplified**: Removed the `ForceFriendlyMessage`
generic parameter - `userFriendlyMessage` is now always required
- **Type safety**: The constructor now requires `{ userFriendlyMessage:
MessageDescriptor }` (no longer optional)
### Updated Files
- **74+ exception classes** updated to provide default user-friendly
messages using Lingui `msg` macro
- Each exception class has a sensible fallback message (e.g., `msg\`An
authentication error occurred.\``)
- Exception classes that had code-specific message maps retain their
behavior
## Benefits
- **Compile-time enforcement**: Forgetting to add a user-friendly
message now causes a TypeScript error
- **Better UX**: End users always see a localized, human-readable error
message
- **Simpler API**: No more boolean generic parameter to think about
## Testing
- `npx nx run twenty-server:typecheck` passes
- `npx nx run twenty-server:lint` passes
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> Enforces `userFriendlyMessage` on `CustomException` and updates all
exception classes to supply localized default messages, with
filters/tests adjusted accordingly.
>
> - **Core**:
> - Enforce required `userFriendlyMessage` in `CustomException` (remove
optional generic; constructor now requires `{ userFriendlyMessage:
MessageDescriptor }`).
> - **Exceptions**:
> - Update ~70+ exception classes to set default localized messages via
Lingui `msg` maps and pass them in constructors (e.g., `AuthException`,
`ObjectMetadataException`, `FieldMetadataException`, etc.).
> - Add fallback messages where needed (e.g., `INTERNAL_SERVER_ERROR` or
domain-specific defaults).
> - **HTTP/GraphQL Filters**:
> - Ensure fallbacks create `UnknownException` with `msg` for
user-friendly text in REST/GraphQL exception filters.
> - **Tests**:
> - Adjust unit tests to pass `userFriendlyMessage` to exceptions.
> - Update Jest snapshots to include `extensions.userFriendlyMessage` or
message objects where applicable.
>
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
221004fdfc0d97b7d152a258b347bf571e70f10e. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
Closes [89](https://github.com/twentyhq/core-team-issues/issues/89)
## Problem
When users attempt to create or update a record with a duplicate value
for a unique field (e.g., duplicate email or domain name), they receive
a generic error message: "This record already exists. Please check your
data and try again." This provides no actionable way to locate and view
the existing conflicting record, forcing users to manually search for
it.
## Solution
This PR enhances duplicate key constraint error handling to
automatically detect the conflicting record and display a "View existing
record" link in the error notification. When clicked, users are
navigated directly to the existing record's detail page.
## Backend Changes
### 1. PostgreSQL Error Parsing
(`parse-postgres-constraint-error.util.ts`)
- Extracts structured information from PostgreSQL `QueryFailedError`
messages.
### 2. Conflicting Record Lookup (`find-conflicting-record.util.ts`)
- Queries the database to find the existing record with the conflicting
value
### 3. Error Handling Orchestration
(`handle-duplicate-key-error.util.ts`)
- Parses PostgreSQL error to extract column name and conflicting value
- Attempts to find the conflicting record
- Enriches `TwentyORMException` with `conflictingRecordId` and
`conflictingObjectNameSingular` if found
### 4. Exception Computation Updates (`compute-twenty-orm-exception.ts`)
- Made function `async` and added optional `entityManager` and
`internalContext` parameters
- Needed to support async database queries for conflicting record lookup
### 5. GraphQL Error Handler
(`twenty-orm-graphql-api-exception-handler.util.ts`)
- **Changes**: Enhanced `DUPLICATE_ENTRY_DETECTED` case to include
`conflictingRecordId` and `conflictingObjectNameSingular` in GraphQL
error extensions
## Frontend Changes
### 1. Error Extraction Utility
(`get-conflicting-record-from-apollo-error.util.ts`)
- Accesses GraphQL error extensions
- Validates that both `conflictingRecordId` and
`conflictingObjectNameSingular` exist and are strings
- Returns `null` if validation fails
### 2. SnackBar Enhancement (`useSnackBar.ts`)
- Extracts conflicting record info from Apollo error
- Constructs URL using `getAppPath` utility
- Adds link object to snackbar options with text "View existing record"
<img width="931" height="858" alt="image"
src="https://github.com/user-attachments/assets/28137dc7-18ab-4ffe-b669-1f2d4ec264d1"
/>
- moves applicationRoleId to application entity
- add new `APPLICATION` FieldActorSource and `APPLICATION`
JwtTokenTypeEnum value
- create a new token with applicationId when executing a function
- when applicationId is in token, check for application.defaultRole
permissions
-use twenty-shared types in `twenty-sdk/application`
- create a new import from generate called "Twenty" that you can use
directly without having to set TWENTY_API_KEY AND TWENTY_API_URL (keep
metadata or core parameter only)
- provide to serverless unique one time BEARER TOKEN to run it
Result
<img width="977" height="566" alt="image"
src="https://github.com/user-attachments/assets/e78428a0-5b13-4975-aa13-58ee3b32450c"
/>
<img width="910" height="596" alt="image"
src="https://github.com/user-attachments/assets/6ec72bf5-7655-4093-a45e-ad269595a324"
/>
<img width="741" height="568" alt="image"
src="https://github.com/user-attachments/assets/7683944c-fd79-4417-8fb2-8e4815cc112f"
/>
## Summary
- Add code interpreter tool that enables AI to execute Python code for
data analysis, CSV processing, and chart generation
- Support for both local (development) and E2B (sandboxed production)
execution drivers
- Real-time streaming of stdout/stderr and generated files
- Frontend components for displaying code execution results with
expandable sections
## Code Quality Improvements
- Extract `getMimeType` to shared utility to reduce code duplication
between drivers
- Fix security issue: escape single quotes/backslashes in E2B driver env
variable injection
- Add `buildExecutionState` helper to reduce duplicated state object
construction
- Add `DEFAULT_CODE_INTERPRETER_TIMEOUT_MS` constant for consistency
- Fix lingui linting warning and TypeScript theme errors in frontend
## Test Plan
- [ ] Test code interpreter with local driver in development
- [ ] Test code interpreter with E2B driver in production environment
- [ ] Verify streaming output displays correctly in chat UI
- [ ] Verify generated files (charts, CSVs) are uploaded and
downloadable
- [ ] Test file upload flow (CSV, Excel) triggers code interpreter
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> Updates generated i18n catalogs for Polish and pseudo-English, adding
strings for code execution/output (code interpreter) and various UI
messages, with minor text adjustments.
>
> - **Localization**:
> - **Generated catalogs**: Refresh `locales/generated/pl-PL.ts` and
`locales/generated/pseudo-en.ts`.
> - Add strings for code execution/output (e.g., code, copy code/output,
running/waiting states, download files, generated files, Python code
execution).
> - Include new UI texts (errors, prompts, menus) and minor text
corrections.
> - No changes to `pt-BR`; other files unchanged functionally.
>
> <sup>Written by [Cursor
Bugbot](https://cursor.com/dashboard?tab=bugbot) for commit
befc13d02c21e5a6647bc1aa6daa2a89f60b7ef8. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
## Context
Following https://github.com/twentyhq/twenty/pull/16399
Now using the new global orm manager everywhere and returning a
GlobalDatasource/WorkspaceDatasource based on a feature flag.
This means we now need to wrap all our ORM calls within
executeInWorkspaceContext callback (at least for now) so the global
datasource can dynamically hydrate its context via the new store (the
global datasource does not store anything related to workspaces as it is
now a unique singleton). If feature flag is off it still uses local data
stored in the workspace datasource.
Changes:
- as we store date in redis as serialized, let's make all flatEntity
dates as string. This requires changing FlatEntity types and making sure
that entity are converted to flatEntity and flatEntity to dtos
## Context
Deprecating TwentyORMManager in favor of TwentyORMGlobalManager
(temporarily, as this will simplify the ultimate goal to later replace
all usages with the new TwentyORMGlobalManagerV2 which will have a
similar signature)
This means this PR had to refactor a bit of code to pass down the
workspaceId when not available directly as it is now a requirement,
meaning we also deprecated scopedWorkspaceContextFactory to have a less
obscure way to fetch the workspaceId and have something more
declarative.
Step 3 will be to update TwentyORMGlobalManager to use a featureFlag
toggling and use the new GlobalWorkspaceOrmManager internally using the
new cache service
Step 4 will be to remove the feature flag and pg_pool patch
First PR to implement application tokens
- add new application role in twenty-server
- move duplicated constants and types to twenty-shared
- will add role configuration utils into twenty-sdk in another PR
## Summary
Consolidates the AI tool provider architecture by creating a single
`ToolProviderService` as the entry point for all tool generation. This
removes multiple intermediate services and simplifies the codebase.
## Changes
### New Architecture
- **`ToolProviderService`**: Single service for all tool generation
with:
- `getTools(spec)` - Get tools by category with permissions
- `getToolByType(type)` - Get specific tool for workflow execution
- **`ToolCategory` enum**: Declarative specification of tool types:
- `DATABASE_CRUD` - Record CRUD operations
- `ACTION` - HTTP requests, email sending, article search
- `WORKFLOW` - Workflow management tools
- `METADATA` - Object/field metadata tools
- `NATIVE_MODEL` - Model-specific tools (e.g., web search)
- **`ToolSpecification` type**: Clean API for requesting tools with
permissions
### Removed
- `AiToolsModule` - No longer needed
- `ToolService` - Logic inlined into ToolProviderService
- `ToolAdapterService` - Logic inlined into ToolProviderService
- `ToolRegistryService` - Logic inlined into ToolProviderService
### Updated
- All consumers (agents, chat, MCP, workflows) now use
`ToolProviderService`
- Test files updated accordingly
## Stats
- **547 insertions, 1146 deletions** (net ~600 lines removed)
- 4 services deleted
- 1 module deleted
## Testing
- [x] Typecheck passes
- [x] Lint passes
Validations :
- relations count (in common api)
- oneToMany relation nested count (in common)
- requested fields count (in gql)
- root resolver count (in gql)
- root resolver duplicates (in gql)
- specific complexity for metadata / nesting count (in gql)
In this PR, we are adding a new command to reset a message channel.
We are also refactoring a bit cursor reset as we had multiple
implementations at different places in the code base
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
## Problem
buildEntityMetadatas in GlobalWorkspaceOrmManager is computationally
expensive and was running on every executeInWorkspaceContext call. This
method uses
TypeORM's EntitySchemaTransformer and EntityMetadataBuilder to build
metadata for all workspace entities (30-50+ objects with many fields
each).
The resulting EntityMetadata[] is not serialisable which means it cannot
be cached in Redis because they contain:
- Circular references
- Functions/methods
- References to the DataSource instance
## Solution
Extended the workspace cache system to support local-only caching, then
created a cache provider for entityMetadatas.
## Implementation details
Updated @WorkspaceCache decorator (workspace-cache.decorator.ts)
- Added localOnly?: boolean option to skip Redis storage for
non-serializable data
Created WorkspaceEntityMetadatasCacheService
- Computes entity metadatas from DB to avoid race condition, this is
acceptable
Simplified GlobalWorkspaceOrmManager
- Now fetches entityMetadatas from cache instead of rebuilding on every
call
Updated Workspace migration runner - the only entry point where metadata
can change
- Now invalidate the new 'entityMetadata' local cache when
shouldIncrementMetadataGraphqlSchemaVersion is true (== field/object
mutations)
# Introduction
Closes https://github.com/twentyhq/core-team-issues/issues/1980
In this PR we migrate the agent from v1 to v2.
## New FlatRoleTargetByAgentIdMaps
Derivated the `flatRoleTargetMaps` to be building a
`flatRoleTargetByAgentIdMaps` to ease retrieving a roleId to associate
to an agent
## Coverage
Added strong coverage on both failing and successful CRU agents
operations
---------
Co-authored-by: Weiko <corentin@twenty.com>
This PR improves the general UX and DX of boards, by modifying the query
effect to only use paged group by queries.
In this PR we implement two more things in the backend for group by
queries :
- Fixed ORDER BY in the PARTITION BY sub-query (this wasn't working
because it was applied in the main query, so it sorted randomly picked
records, which was a correct sort on an incorrect dataset returned by
the sub-query)
- Added offset paging in PARTITION BY
Miscellaneous, various bug fixes and improvements along the way :
- Throttled loading of cards to avoid React freeze
- Handling of drag & drop
- Handling of create / delete / update
- Reworked skeleton (the library slows down a lot with hundreds of
skeleton for a spinning effect that is hardly noticed)
- Fixed refetch of aggregate queries (I included the new group by
aggregates query we use in the existing refetch mechanism)
- Re-trigger queries on filters and sorts changes
- Unselect all record ids when deleting / restoring / detroying
- Fetch only groups that still have records to lighten the group by
query.
# What remains to be done
This is still a naïve fetch more implementation that will work for a few
fetch more rounds, but if you scroll and load say 200 cards per column
on a board, React will re-render all 200 cards of each column each time.
We would probably need to virtualize the board with paged queries as we
did for the table, this could be done after this PR but seems less
urgent.
What's nice is that this new query pattern is well designed for
virtualization also, drawing from our experience with table
virtualization, and adapted to a multi-column request pattern, like a
2:2 matrix of records, for our boards.
So the remaining work would be to design a UI solution for virtualizing
this matrix of records, which could be quite different from our table
virtualization mechanism.
## Context
We've recently introduced a new workspace cache service which now acts
as a cache access and local storage for all workspace related data,
deprecating the individual specific services.
- Better performance through multiple caching/fetching strategies
- Consistent data access patterns across the codebase
- Reduced redis queries through MGET/MSET/PIPELINE with multiple cache
keys
## Context
Deprecating the old objectMetadataMap type in favour of split flat
entities to match with our new caching.
In the long run, trying to achieve:
- Better performance through caching
- Consistent data access patterns across the codebase
- Reduced database queries
Now that everything is based on flat entities, which are cached, we can
finish the refactoring of workspace context cache which should already
improve performances.
Then the last step will be to consume that new cache in the new global
datasource to get rid of the many workspace datasources stored in the
server
## Summary
This PR introduces a comprehensive agent evaluation system and refactors
the AI module structure for better organization.
## Key Changes
### 🎯 Agent Evaluation System
- Added **Agent Turn Evaluation** entities, DTOs, and database schema
- New GraphQL mutations: `evaluateAgentTurn` and `runEvaluationInput`
- Added `evaluationInputs` field to Agent entity for storing test inputs
- New `AgentTurnGraderService` for automatic turn evaluation
- Added evaluation UI with new **Evals** and **Logs** tabs in agent
detail pages
### 🏗️ Entity & Module Refactoring
- Renamed `AgentChatMessage` → `AgentMessage` for clarity
- Consolidated chat entities: `AgentMessage`, `AgentTurn`, and
`AgentChatThread`
- Reorganized AI modules under `ai/` subdirectory structure
- Updated imports across codebase to reflect new module paths
### 🤖 New Agents & Roles
- Added **Dashboard Builder Agent** for dashboard creation and
management
- Added **Dashboard Manager Role** with appropriate permissions
- Updated role permissions to be more granular (users vs agents vs API
keys)
### 🔐 Permission System Updates
- Added `HTTP_REQUEST_TOOL` permission flag
- Updated Workflow Manager role permissions (restricted tool access)
- Enhanced permission flag types to differentiate between user/agent/API
key contexts
- Added `isRelevantForAgents`, `isRelevantForApiKeys`,
`isRelevantForUsers` to permission flags
### 📨 Message Role Enhancement
- Added `system` role to `AgentMessageRole` enum (alongside
user/assistant)
- Updated message handling to support system prompts
### 🎨 UI/UX Improvements
- New tabs in agent detail: **Evals** and **Logs**
- Added turn detail page: `/ai/agents/:agentId/turns/:turnId`
- Fixed text overflow in `SettingsListItemCardContent`
- Updated role applicability labels ("Assignable to Workspace Members")
### 🛠️ Technical Improvements
- Fixed Zod schema validation for UUID and Date fields (use string
validators)
- Updated `ToolRegistryService` to properly register HTTP tool with
permission flag
- Enhanced error handling in agent execution services
- Updated database migrations for new entity schema
## Database Migrations
- `1764210000000-add-system-role-to-agent-message.ts`
- `1764220000000-add-evaluation-inputs-to-agent.ts`
- `1764200000000-add-agent-turn-evaluation.ts`
- `1764100000000-refactor-agent-chat-entities.ts`
## Testing
- [ ] Agent evaluation flow tested
- [ ] Dashboard Builder agent tested
- [ ] Permission system validated
- [ ] UI tabs and navigation tested
- [ ] Database migrations run successfully
## Breaking Changes
⚠️ **Entity Rename**: `AgentChatMessage` renamed to `AgentMessage` -
GraphQL queries need updating
## Related Issues
<!-- Link any related issues here -->
## Screenshots
<!-- Add screenshots if applicable -->
## Overview
This PR replaces the dynamic agent handoff system with a more
predictable planning-based router that decides upfront how to handle
multi-agent coordination.
## Major Changes
### 🔄 Architecture Shift: Handoffs → Planning
**Removed:**
- `AgentHandoffEntity` and handoff tracking system
- `AgentHandoffService` and `AgentHandoffExecutorService`
- Dynamic agent-to-agent transfers during execution
- Handoff tool generation and description templates
**Added:**
- `AiRouterService` with two strategies: `simple` (single agent) and
`planned` (multi-agent)
- `AgentPlanExecutorService` for executing multi-step plans
- Plan validation (cycle detection, dependency resolution)
- `UnifiedRouterResult` type with discriminated union
### 🤖 New Standard Agents
Added two new specialized agents:
- **Researcher Agent**: Web search, fact-finding, competitive
intelligence
- **Code Agent**: TypeScript function generation for serverless
workflows
### 🏗️ Router Refactoring (Latest)
Split router responsibilities into focused services:
- `AiRouterStrategyDeciderService`: Decides simple vs planned strategy
- `AiRouterPlanGeneratorService`: Generates and validates execution
plans
- `AiRouterService`: Coordinates between services (reduced from 426→275
lines)
### ⚙️ Configuration Improvements
- Added `outputStrategy` to agent definitions (`direct` vs `synthesize`)
- Removed hardcoded special cases for workflow-builder
- Added `plannerModel` field to workspace entity
- Increased `MAX_STEPS` from 10 to 25 for complex workflows
### 📝 Agent Prompt Refinements
Significantly simplified prompts for better clarity:
- Workflow Builder: 51→36 lines
- Helper: 49→28 lines
- Data Manipulator: Enhanced with sorting guidance
### 🔍 Enhanced Debugging
- Plan reasoning and step count in data message parts
- Router debug info with token usage tracking
- Better logging throughout execution pipeline
## Benefits
1. **Simpler Mental Model**: Router decides upfront vs dynamic transfers
2. **Better Predictability**: Users see the plan before execution
3. **Cleaner Architecture**: SRP with focused services
4. **Configuration Over Code**: Agent behavior via config, not hardcoded
logic
5. **Plan Validation**: Catches invalid dependencies and cycles
## Migration Notes
- Database migration removes `agentHandoff` table
- Adds `plannerModel` column to workspace table
- No API breaking changes (agent endpoints unchanged)
## Testing
- Integration tests updated to remove handoff dependencies
- Agent tool test utilities simplified
- Plan validation covered by new logic
## Next Steps (Future PRs)
- Parallel execution of independent plan steps
- Dynamic re-planning based on results
- Plan caching for common routing patterns
- Error recovery strategies in plan executor
## Context
Deprecating legacy ObjectMetadata from cache in favor of flat entities.
Introducing utils to build byName/byNameSingular/byNamePlural in
isolated cases
## Next
- I had to introduce a util to build from flat to legacy
objectMetadataMaps, we should instead use flat maps directly when needed
(datasource, schema generation, etc)
- Deprecate metadata version in the cache
- Use the new cache strategy for flat entities with permissions and
feature flags and inject in the global datasource context
closes https://github.com/twentyhq/core-team-issues/issues/1629
To do before requesting review :
- filter update
Migration to come in an other PR
Strat :
1/ Null transformation
- [x] Transform NULL equivalent value to NULL in field validation in
common api - pre-query - with feature flag
- [ ] Same logic in ORM (Not done, complex to handle feature flag here)
- [x] Transform NULL value to equivalent in data formatting in ORM -
post-query
2/ Migration (in other PR) for fieldMetadata not nullable with default
defaultValue (empty string, ...)
- [ ] Remove NOT NULL db constraint
- [ ] Update record value to NULL
- [ ] Update field metadata : isNullable:true
- [ ] Update uniqueIndex whereClause (also for standard uniqueIndex)
- [ ] Activate feature flag
3/ Update metadata creation
- [x] No more default default value
- [x] Update standard field nullability
- [x] Remove index default whereClause for standard field
4/ Update filter
- [x] When filtering on NULL or empty string, be sure all records are
returned (the one with NULL + the one with "")
5/ Test
- [ ] Strat. to do
Weiko