## 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
# Introduction
Two things:
- Enforcing non nullable workspace custom application Id for any
workspace
- Fixing front non editable data models following
https://github.com/twentyhq/twenty/pull/15911 that associate any custom
entities to an applicationId. The front was putting everything as
readonly when under an app ( we will have to handle the twenty standard
application in the future too )
## Fallback
### Migration
The non nullable migration will fail when released, that's why it's
being swallowed and re-run in an upgrade command post workspace custom
application creation for those that miss one. Allowing the migration to
pass in the end
The typeorm migration still need to exists for any new workspaces
### GetCurrentUser
In order to dynamically display isReadOnly in data model settings we're
fetching the workspaceCustomApplicationId through the `getCurrentUser`
If not fallback this endpoint would throw until we're handling existing
workspaces that do not have a custom workspace application
The fallback should be removed post release
# Introduction
Cleaner and fewer scope version of
https://github.com/twentyhq/twenty/pull/15745 ( removed sync-metadata
hack through, too ambitious migration and upgrade )
Please note that this PR won't have any interaction with the existing
sync-metadata
Which mean that the sync metadata does not update the standard entities
applicationId and universalIdentifier, and it won't we will deprecate it
on favor of a workspace migration aka twenty-standard app installation
## API Metadata
Any operation going through the api metadata nows automatically scope
the related entity to the workspace custom application instance. (
optionally passing an applicationId to allow current hacky implem of app
sync service )
We need to either ignore the tests or remove the cli status check from
the blocking status badges for a PR to be merged
## New workspace
Already handled in previous
https://github.com/twentyhq/twenty/pull/15625, when a workspace is
created it gets created a twenty standard and custom workspace instance
All his views and permissions will be prefilled to the its twenty
standard app instance with a specific universalIdentifier
## New universalIdentifier
At the contrary as before with standardIds, universalIdentifier are
unique for a given workspace
This means that createdAt field of both object company and opportunity
will have a unique universalIdentifier whereas they share the same
standardId
## FlatApplication
Introduced the flatApplication and cache. Will migrate existing
`MetadataName` to be `SyncableMetadataName` in a following PR
## What's next
Next we will describe a twenty standard app configuration as json that
will be used to generate a workspace migration that will be run instead
of the sync metadata, in a nutshell we aim to deprecated the sync
metadata
So we can standardize any entity to have a non nullable applicationId
and universalIdentifier
## Upgrade command
Introduced an upgrade command that will create a custom workspace
instance for any workspace that do not have one in order to align with
the new behavior when creating a new workspace
# Introduction
Remove the v2 feature flag for view-field field-metadata and
object-metadata metadata entities
## Some details
- Disabled nestjs-query for object metadata creation and explicitly
calling it
- removed all v1 integration tests files
## Remarks
Not remove v2 referencing in both filenaming right now will handle that
globally later
## Breaking change
Due to object metadata resolver createOne standardization had to rename
the input from `CreateObjectInput` to `CreateOneObjectInput`
## Context
inferDeletionFromEntities only accepts a set of keys for each entities
that needs deletion. This could be error-prone if tmr we want to add a
new side effect and forget to add the entity when in practice you want
to delete all entities that are in the fromToAllFlatEntityMaps (this is
the case for applications for example)
# Introduction
related to https://github.com/twentyhq/core-team-issues/issues/1833
In this PR we're starting the sync-metadata and standardIds deprecation
by introducing `twenty-standard` application that will regroup every
standard object such as company and opportunities. But also the
`custom-workspace-application` which is an app created at the same time
as a workspace and that will regroup everything configure within the
workspace ( custom objects fields etc )
## What's done
On both new workspace and seeded workspace creation:
- Creating a custom workspace app
- Creating a twenty standard app
- Refactored the seed core schema and workspace creation to be run
within a transaction in order to handle circular dependency foreignkey
requirements ( which is deferred for app toward workspace )
- Updated workspace entity to have a custom workspace relation (
nullable for the moment until we implem an upgrade command to handle
retro comp )
- Integration testing on user, workspace creation deletion and expected
default apps creation
- ~~Soft deleted user on `deleteUser`~~ Done by marie and rebased on it
## What's next
- Update seeder to propagate the `twenty-standard` workspace
`applicationId` to every standard synchronized entities ( cheap and fast
iteration through the about to be deprecated sync-metadata as an easy
way to synchronize standards metadata entities ).
- Update seeder to propagate the `custom-workspace-application`
workspace `applicationId` to anything custom ( `pets` and `rockets` )
- Prepend `custom-workspace-application` `applicationId` to every
metadata API operations ( create a specific cache etc )
- Upgrade command on all existing workspace to create a custom app and
associate its applicationId to any existing custom entities
- Make `universalIdentifier` and `applicationId` required for any
syncable entity
## Problem
The ESLint rule `graphql-resolvers-should-be-guarded` introduced in
#15392 was failing on main because the guard `SettingsPermissionsGuard`
had inconsistent naming.
## Root Cause
The guard was named `SettingsPermissionsGuard` (with an 's') which was
inconsistent with other permission guards:
- ✅ `CustomPermissionGuard`
- ✅ `NoPermissionGuard`
- ✅ `ImpersonatePermissionGuard`
- ❌ `SettingsPermissionsGuard` (inconsistent!)
The ESLint rule checks if guard names end with `PermissionGuard`, but
`SettingsPermissionsGuard` ends with `sGuard`, so it wasn't recognized
as a permission guard.
## Solution
Renamed the guard to be consistent with the naming convention:
1. ✅ Renamed file: `settings-permissions.guard.ts` →
`settings-permission.guard.ts`
2. ✅ Renamed export: `SettingsPermissionsGuard` →
`SettingsPermissionGuard`
3. ✅ Renamed internal class: `SettingsPermissionsMixin` →
`SettingsPermissionMixin`
4. ✅ Updated all 122 references across 44 files in the codebase
5. ✅ Renamed test file: `settings-permissions.guard.spec.ts` →
`settings-permission.guard.spec.ts`
## Testing
- ✅ `npx nx run twenty-server:lint` passes
- ✅ `npx nx run twenty-server:typecheck` passes
- ✅ No references to the old name remain in the codebase
- ✅ All previously failing resolver files now pass ESLint validation
## Related
Fixes issues introduced in #15392
## Overview
This PR strengthens our permission system by introducing more granular
role-based access control across the platform.
## Changes
### New Permissions Added
- **Applications** - Control who can install and manage applications
- **Layouts** - Control who can customize page layouts and UI structure
- **AI** - Control access to AI features and agents
- **Upload File** - Separate permission for file uploads
- **Download File** - Separate permission for file downloads (frontend
visibility)
### Security Enhancements
- Implemented whitelist-based validation for workspace field updates
- Added explicit permission guards to core entity resolvers
- Enhanced ESLint rule to enforce permission checks on all mutations
- Created `CustomPermissionGuard` and `NoPermissionGuard` for better
code documentation
### Affected Components
- Core entity resolvers: webhooks, files, domains, applications,
layouts, postgres credentials
- Workspace update mutations now use whitelist validation
- Settings UI updated with new permission controls
### Developer Experience
- ESLint now catches missing permission guards during development
- Explicit guard markers make permission requirements clear in code
review
- Comprehensive test coverage for new permission logic
## Testing
- ✅ All TypeScript type checks pass
- ✅ ESLint validation passes
- ✅ New permission guards properly enforced
- ✅ Frontend UI displays new permissions correctly
## Migration Notes
Existing workspaces will need to assign the new permissions to roles as
needed. By default, all new permissions are set to `false` for non-admin
roles.
As title
- adds decorators in twenty-sdk
- update twenty-cli load-manifest to it gets @FieldMetadata infos +
testing
- update twenty-server so it CRUD fields properly, using
universalIdentifier
- Fix UI so we can update managed objects records
- move FieldMetadata items from twenty-server to twenty-shared
Several fixes after discussing with @BOHEUS
- set applicationManifest env key optional
- fix server local serverless function logging (introduces a new env
variable `SERVERLESS_LOGS_ENABLED` defaulting to false)
## Context
- All flatEntity should extend SyncableEntity
- SyncableEntity should now have applicationId and application relation
- Fix syncApp deletion, should now properly use migration v2 to delete
syncable entities
# Introduction
This PR introduces a huge type refactor that will leverage dynamic intra
entity optimistic flat maps update in the future and also a more
granular cache invalidation enhancing performances
close https://github.com/twentyhq/core-team-issues/issues/1717
close https://github.com/twentyhq/core-team-issues/issues/1716
close https://github.com/twentyhq/core-team-issues/issues/1643
## What's done
### Comparators centralization
Comparator is now done through global configuration as const for each
metadata names
Thanks to
Note:
Definition of standard is evolving, standard is now scoped to an app.
Meaning that a manifest should be able to update its own standards
objects but on other app standards ones ? Each synchronizable entities
will have a standardOverrides ?
## Typing refactor
### `AllFlatEntityTypesByMetadataName`
**Single source of truth for the complete type ecosystem**, mapping each
metadata name to its entity types, flat entities, and migration actions:
```typescript
export type AllFlatEntityTypesByMetadataName = {
fieldMetadata: {
actions: { created: CreateFieldAction; updated: UpdateFieldAction; deleted: DeleteFieldAction; };
flatEntity: FlatFieldMetadata;
entity: FieldMetadataEntity;
};
objectMetadata: { /* ... */ };
// ... all 10 metadata types
};
```
### `ALL_METADATA_NAME_MANY_TO_ONE_RELATIONS`
**Explicitly declares database relationships** between entities with
compile-time validation:
```typescript
export const ALL_METADATA_NAME_MANY_TO_ONE_RELATIONS = {
viewField: {
view: 'viewId',
fieldMetadata: 'fieldMetadataId',
},
cronTrigger: {
serverlessFunction: 'serverlessFunctionId',
},
// ... all relations
} as const satisfies MetadataNameAndRelations;
```
### `ALL_FLAT_ENTITY_CONFIGURATION`
**Centralizes comparison and serialization logic** for each metadata
type:
```typescript
export const ALL_FLAT_ENTITY_CONFIGURATION = {
fieldMetadata: {
propertiesToCompare: ['name', 'type', 'label', 'defaultValue', /* ... */],
propertiesToStringify: ['options', 'settings', 'defaultValue'],
},
objectMetadata: {
propertiesToCompare: ['nameSingular', 'namePlural', 'isActive', /* ... */],
propertiesToStringify: [],
},
// ... all metadata types
} as const satisfies AllFlatEntityConfiguration;
```
## Combined Impact
These three configurations work together to create a **strongly-typed,
centrally-managed metadata system**:
1. **`AllFlatEntityTypesByMetadataName`** defines *what exists*
2. **`ALL_METADATA_NAME_MANY_TO_ONE_RELATIONS`** defines *how they
relate*
3. **`ALL_FLAT_ENTITY_CONFIGURATION`** defines *how to compare and
serialize them*
**Result:** Builders and validators become thin wrappers around
type-safe, configuration-driven logic instead of containing scattered,
error-prone manual implementations.
## What's next
### StandardOverrides standardization
Every metadata entity can be a standard one for a workspace if it's an
installed app, which means it might not expose the whole entity api to
be editable through an import dynamically
The standard overrides logic should not be applied to Fields and Objects
but to every entities
At the moment we have a logic of `EDITABLE_PROPERTIES` through the api,
and also `STANDARD_OVERRIDEDABLE_PROPERTIES`
This should be configuration centered like `propertiesToCompare` and
`propertiesToStringify`.
Scoping this PR to two last for the moment. As update dispatch to
standardOverrides could be considered as a side effect prefer waiting to
start the side effect refactor
### Granular Optimistic deprecation
With this new grain at runtime we will be able to add a flat entity and
dispatch its addition to related flat maps, so we don't have to describe
an optimistic method for each flat entity operations
See `addFlatEntityToFlatEntityAndRelatedEntityMapsOrThrow`
Note: Still in wip and included in this PR but about to create a new one
to integrate these utils and remove existing methods
### ValidateBuildAndRun dynamic args typed defintion
We should restrain the devxp to send expected flat maps entity as at
least from to or dependency as we now have the grain both a type lvl and
runtime to do so
It should not be possible in the devxp to forgot adding the views to the
v2 builder when passing the view field anymore ( that would lead to
permanent validation error in view field integrity checks )
## Conclusion
Thanks for reading and reviewing !
Any suggestions are more than welcomed ! ( same as for questions too ! )
App deletion currently deletes all related entities through postgres
cascade delete meaning we never run actual migrations ourself and never
trigger cache invalidation (part of migration engine)
When starting the app on a fresh database reset the cache would be
filled with empty flat field metadata maps
Because the reset command hack through the repository directly in order
to create views and stuff
implemented an iso flush as the one existing initially
added it to a workspace deletion tambien
```
{
byId: {],
universalIdById: {}
}
```
## Context
Add routes to migration V2
- Resolvers
- Service v2
- Builder
- Validator
- Action runner
Next PR: Add to twenty-cli to sync routes with serverless
# Introduction
Extracting the legacy fields build and dispatch out of the object one to
follow the generic flat entity build, also update caches entries for
object
## Main tasks
- flat field map cache
- flat field builder
- refactored the dispatch matrix to return flat entity maps instead of
flat entity arrays
- orchestrator aggregator
- removing legacy code
- making universal identifier aka standardId of standard field for
custom object deterministically dynamic
- perfs debug logs for v2
## TODO
- [x] Refactor the generic entity builder to be dependency flat maps in
order to main foreign keys list in flat parent
- [x] Refactor the flat object metadata to contain the array of related
fields and avoid costy find object fields
- [ ] Improve the create field handler to handle multiple field at once
- [ ] Refactor the dispatch to embbed the comparison
- [ ] Improve perf by extracting from elements out of existing
- [ ] Fix the labelIdentifierId validators on object before field
creation ( integ tests are in failing mode )
## Debug logs snippet
```ts
[EntityBuilder fieldMetadata] matrix computation: 0.027ms
[EntityBuilder fieldMetadata] creation validation: 0.001ms
[EntityBuilder fieldMetadata] deletion validation: 0.293ms
[EntityBuilder fieldMetadata] update validation: 0.006ms
[EntityBuilder fieldMetadata] entity processing: 0.363ms
[EntityBuilder fieldMetadata] validateAndBuild: 0.455ms
[EntityBuilder index] matrix computation: 0.005ms
[EntityBuilder index] creation validation: 0.001ms
[EntityBuilder index] deletion validation: 0.146ms
[EntityBuilder index] update validation: 0.004ms
[EntityBuilder index] entity processing: 0.199ms
[EntityBuilder index] validateAndBuild: 0.228ms
[Runner] Initial cache retrieval: 0.549ms
[BaseWorkspaceMigrationRunnerActionHandlerService] delete_index executeForWorkspaceSchema: 11.665ms
[BaseWorkspaceMigrationRunnerActionHandlerService] delete_index executeForMetadata: 12.864ms
[BaseWorkspaceMigrationRunnerActionHandlerService] delete_field executeForWorkspaceSchema: 1.476ms
[BaseWorkspaceMigrationRunnerActionHandlerService] delete_field executeForMetadata: 6.816ms
[BaseWorkspaceMigrationRunnerActionHandlerService] delete_field executeForWorkspaceSchema: 0.062ms
[BaseWorkspaceMigrationRunnerActionHandlerService] delete_field executeForMetadata: 0.889ms
[Runner] Transaction execution: 23.434ms
[Runner] Cache invalidation: 316.662ms
[Runner] Total execution: 340.767ms
```
As you can see cache invalidation is way to long, we could replace the
cache by the optimistic in the end
twenty-cli serverless triggers follow up. Fixes:
- eventName don't support wildcard
- universalIdentifier not used to create or update trigger : update does
not work properly (does deletion then creation)
- add a base project in twenty-cli that is copied when creating a new
app
- add serverlessFunction schema in twenty-cli
- add trigger schema in twenty-cli
- update serverless function code save and get
- sync serverless function
Félix Malfait