⚠️ **AI-generated PR — not ready for review** ⚠️
cc @FelixMalfait
---
## Changes
### System prompt improvements
- Explicit skill-before-tools workflow to prevent the model from calling
tools without loading the matching skill first
- Data efficiency guidance (default small limits, use filters)
- Pluralized `load_skill` → `load_skills` for consistency with
`load_tools`
### Token usage reduction
- Output serialization layer: strips null/undefined/empty values from
tool results
- Lowered default `find_*` limit from 100 → 10, max from 1000 → 100
### System object tool generation
- System objects (calendar events, messages, etc.) now generate AI tools
- Only workflow-related and favorite-related objects are excluded
### Context window display fix
- **Bug**: UI compared cumulative tokens (sum of all turns) against
single-request context window → showed 100% after a few turns
- **Fix**: Track `conversationSize` (last step's `inputTokens`) which
represents the actual conversation history size sent to the model
- New `conversationSize` column on thread entity with migration
### Workspace AI instructions
- Support for custom workspace-level AI instructions
---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
## Summary
- Deletes `use-throttler.ts` — an envelop plugin that was never
registered in the GraphQL config plugin chain (dead code since
introduction)
- Removes the `graphql-rate-limit` dependency from `package.json` (only
consumer was the deleted file)
API rate limiting continues to work via `ThrottlerService` in the query
runner layer (for API key auth). Global coverage should be handled at
the infrastructure level (Cloudflare).
## Test plan
- [x] Confirmed no imports of `useThrottler`, `ThrottlerPluginOptions`,
or `UnauthenticatedError` from this file exist anywhere in the codebase
- [x] `graphql-rate-limit` has no other consumers
Made with [Cursor](https://cursor.com)
Co-authored-by: Cursor <cursoragent@cursor.com>
## Summary
This PR implements credit rollover functionality for billing, allowing
unused credits from one billing period to carry over to the next (capped
at the current period's subscription tier cap).
## Changes
### New Services
- **StripeCreditGrantService**: Interacts with Stripe's Billing Credits
API to create credit grants, retrieve customer credit balances, and void
grants
- **BillingCreditRolloverService**: Contains the rollover logic -
calculates unused credits and creates new grants for the next period
- **BillingWebhookCreditGrantService**: Handles
`billing.credit_grant.created` and `billing.credit_grant.updated`
webhooks to update billing alerts
### Modified Services
- **StripeBillingAlertService**: Updated to include credit balance when
calculating usage threshold alerts
- **BillingUsageService**: Returns rollover credits to the frontend for
display
- **BillingSubscriptionService**: Queries credit balance when creating
billing alerts
- **BillingWebhookInvoiceService**: Triggers rollover processing on
`invoice.finalized` webhook
### Frontend
- Updated `SettingsBillingCreditsSection` to display base credits,
rollover credits, and total available
- Updated GraphQL query to fetch new `rolloverCredits` and
`totalGrantedCredits` fields
### Stripe SDK Upgrade
- Upgraded from v17.3.1 to v19.3.1 to get proper types for
billing.credit_grant events
- Fixed breaking changes: invoice.subscription path, subscription period
fields location, removed properties
## How it works
1. When `invoice.finalized` webhook is received for
`subscription_cycle`, the system:
- Calculates usage from the previous period
- Determines unused credits (tier cap - usage)
- Caps rollover at current tier cap
- Creates a Stripe credit grant with expiration at end of new period
2. When credit grants are created/updated/voided:
- Billing alerts are recreated with the updated credit balance
3. The UI displays:
- Base credits (from subscription tier)
- Rollover credits (from previous periods)
- Total available credits
## Edge Cases Handled
- Credit grant voided: `billing.credit_grant.updated` webhook triggers
alert update
- Credit grant expired: Stripe's `creditBalanceSummary` API excludes
expired grants
- No unused credits: Rollover service skips grant creation
- Customer ID as object: Controller extracts `.id` from expanded
customer
- Batch Gmail API calls using `googleapis-batcher` for folder processing
- Add concurrency limit for Microsoft Graph folder processing
- Skip IMAP folder sync when no new messages (checks UIDVALIDITY/MODSEQ)
- Refactored `syncMessageFolders` to return folder state directly,
avoiding extra DB round-trips
- Refactored `processPendingFolderActions` to reuse state instead of
querying DB again
- Add unique index on message folders entity
---------
Co-authored-by: Charles Bochet <charles@twenty.com>
## 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
befc13d02c. This will update automatically
on new commits. Configure
[here](https://cursor.com/dashboard?tab=bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
Resolves [Dependabot Alert
323](https://github.com/twentyhq/twenty/security/dependabot/323),
[Dependabot Alert
324](https://github.com/twentyhq/twenty/security/dependabot/324) and
[Dependabot Alert
325](https://github.com/twentyhq/twenty/security/dependabot/325).
It updates Sentry's packages on the server from 10.21.0 to 10.27.0.
I also moved @sentry/react to twenty-front package.json and updated the
version from 9.26.0 to 10.27.0 - no breaking changes were introduced in
the major upgrade in regards to the API exposed by the dependency.
Since @sentry/profiling-node was redundant in the root package.json, I
removed it - twenty-server has it already and is the only package
dependent on @sentry/profiling-node.
Resolves [Dependabot Alert
74](https://github.com/twentyhq/twenty/security/dependabot/74).
Upgraded `graphql-upload` from `13.0.0` to `16.0.2`. Type exports
changed. API remains the same.
Tested the following upload flows:
- profile picture
- workspace logo
- rich text editor (image, video, file)
- record profile picture
- file associated to record.
They all work as intended, nothing breaks.
There was an extra definition of @nx/js inside twenty-server, which was
not recognized during upgrade by the CLI. This caused two versions of
@nx/js in the repo - 22.0.3 from root and 21.3.11 from the twenty-server
package.
Removed the one inside twenty-server to maintain a single source of
truth.
Some Glob CLI related alerts were generated last night. I believe
they're safe to dismiss since I do not expect us to use the Glob CLI in
production environment, and those alerts do not impact the API, but
still updating the dependency version just in case.
Not sure if this would resolve all/any of the alerts since glob is a
dependency for many other dependencies, so a good number of dependency
variants are pulled in.
The alert that confirms it's just a CLI related vulnerability:
[Dependabot Alert
307](https://github.com/twentyhq/twenty/security/dependabot/307)
Resolves [Dependabot Alert
301](https://github.com/twentyhq/twenty/security/dependabot/301).
Locked the version of "ai" at 5.0.52 in order to stop it from bumping to
5.0.93 directly when the ^ is introduced since it breaks the code across
multiple files.
## Introduction
Since we've moved some class validator instances from twenty-server to
twenty-shared tests are red because they do not know how to parse
decorators declarations
We've fixed this by explicitly installing `class-validator` in
`twenty-shared` and configuring jest swc accordingly
## Twenty-server class validator patch
I don't even know if that's something we need anymore
Seems to be a patch either fixing or introducing credit card and phone
number validation
Would prefer discussing the need or not to either before merging this as
it could introduce regression at runtime:
- Centralize the patch to be consumed in both `twenty-server` and
`twenty-shared`
- Remove the patch
We should also document every patch motivations we do as it's quite
though to iterate over a such huge one
Resolves [Dependabot Alert
289](https://github.com/twentyhq/twenty/security/dependabot/289) and a
couple other alerts.
Removed types for `imapflow` since the package ships them internally
now. `yarn.lock` has major changes due to an upgraded AWS SDK
`@aws-sdk/client-sesv2` which is used by Nodemailer 7.
- No breaking changes were introduced in imapflow and mailparser.
- Nodemailer's breaking change was dropping the legacy SES transport; we
already use the SMTP transport + our own AWS SES client, so nothing else
needs changing.
This PR implements Sentry's AI agent monitoring by:
- Configuring vercelAIIntegration with recordInputs and recordOutputs
options
- Adding sendDefaultPii to Sentry.init() for better debugging
- Creating a shared AI_TELEMETRY_CONFIG constant to DRY up the telemetry
configuration
- Adding experimental_telemetry to all AI SDK calls (generateText,
generateObject, streamText)
All AI operations are now fully monitored in Sentry with complete
input/output recording for debugging and performance analysis.
Note: Currently on Sentry v9.26.0, which is compatible with this
implementation. No breaking changes from the v9-to-v10 migration guide
were found in the codebase.
Instead of declaring packages at the workplace-level package.json,
moving them into their relevant package-level package.json file.
`twenty-front`, `twenty-server` and `twenty-emails` continue to build
and work fine because of hoisting, but the dependencies now follow the
internal strategy of declaring at the package-level, plus give us a
single source of truth to updating package versions.
`twenty-front` and `twenty-emails` only use `@react-email/components`
while `twenty-server` only depends on `@react-email/render`.
Fixes [Dependabot Alert
85](https://github.com/twentyhq/twenty/security/dependabot/85) -
prototype pollution in lodash.
Added a shared pick helper (with unit tests) in twenty-shared and
refactored front-end/server code to import { pick } from the shared
barrel instead of lodash.pick.
---------
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
Co-authored-by: martmull <martmull@hotmail.fr>
- Added 'Cron will be triggered at UTC time' notice below trigger
interval dropdown
- Positioned correctly between dropdown and expression field to match
design
- Only shows when Custom CRON option is selected
---------
Co-authored-by: Félix Malfait <felix@twenty.com>
Co-authored-by: Félix Malfait <felix.malfait@gmail.com>
This Pr
- adds a `route` table to the core schema
- adds a controller to trigger route
For now, we need to add a `/s/<workspace_id>` prefix to all routes
We plan to create a custom domain table in order to let the users create
subdomains for their workspace, and so to link their routes to a
subdomain. Thank to that, we will be able to identify workspace_id from
domain name, and the prefix could be `/s`. If we create a native
dedicated subdomain for routes for each workspaces, the prefix could be
completely removed!
Here the follow up ticket to do that ->
https://github.com/twentyhq/twenty/issues/14240
- Move dev-only types to devDependencies
- Move frontend-only deps from root to twenty-front
- Add website-only deps to twenty-website
- Fix react-phone-number-input patch path
CI should validate.
Testing a different approach to fix broken buildPackageJson on server
build
How i have tested:
A. Local contributor setup
- run yarn
- build server
- run yarn workspace focus
- run server on dist
B. self-host
- docker build
Note: I think the dependencies I have added are suboptimized as the
image went from 2GB to 3.5GB. We might need to be more accurate
Unless I'm mistaken the project does not run with node `24.0.0`
Switching to node `24.5.0` ( as defined in vscode node runtime
requirements in https://github.com/twentyhq/twenty/pull/13730 ) seems to
fix the issue
```ts
Successfully compiled: 2897 files with swc (188.32ms)
(node:77006) [DEP0190] DeprecationWarning: Passing args to a child process with shell option true can lead to security vulnerabilities, as the arguments are not escaped, only concatenated.
(Use `node --trace-deprecation ...` to show where the warning was created)
Watching for file changes.
/Users/paulrastoin/ws/twenty/node_modules/buffer-equal-constant-time/index.js:37
var origSlowBufEqual = SlowBuffer.prototype.equal;
^
TypeError: Cannot read properties of undefined (reading 'prototype')
```
Updating engines so local constraint suggest a functional node version
