apps/api/src/app/agents/usecases/get-agent/get-agent.usecase.ts (1)

13-20: Consider narrowing select to response fields only.

Line 13-20 is correctly scoped, but using '*' loads full documents. If toAgentResponse needs a subset, prefer array select to reduce DB/network payload.

As per coding guidelines: use array syntax as the default for select, with '*' as explicit full-entity opt-in.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/usecases/get-agent/get-agent.usecase.ts` around lines
13 - 20, The findOne call on agentRepository is currently using '*' which
returns the whole document; change it to an explicit array select containing
only the fields required by toAgentResponse (e.g., the id/identifier, name,
status, config, etc. used in ToAgentResponse) so you limit DB/network payload;
update the select argument in the agentRepository.findOne(...) call to that
array form rather than '*' and ensure toAgentResponse still receives all
required fields.

apps/api/src/app/agents/usecases/list-agents/list-agents.usecase.ts (1)

13-20: Prefer explicit field selection for list reads when possible.

Line 13-20 is functionally correct, but list endpoints are good candidates for array select to avoid full-document reads if the mapper uses only a subset.

As per coding guidelines: default to array syntax for select, and use '*' only when full entity retrieval is needed.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/usecases/list-agents/list-agents.usecase.ts` around
lines 13 - 20, The list use case currently calls this.agentRepository.find(...,
'*', ...) which reads full documents; change the select argument to an explicit
array of fields used by the mapper (e.g.,
['_id','name','status','_environmentId','_organizationId','createdAt']) so list
reads only necessary columns; update the this.agentRepository.find call in
list-agents.usecase.ts to pass that array instead of '*' and ensure any mapped
properties are included in the array.

apps/api/src/app/agents/dtos/create-agent-request.dto.ts (1)

5-18: Add length and format guards for name/identifier.

Current validation accepts any non-empty string. Consider adding sane max lengths and a URL-safe identifier pattern to prevent invalid or hard-to-route identifiers.

Proposed hardening

-import { IsNotEmpty, IsOptional, IsString } from 'class-validator';
+import { IsNotEmpty, IsOptional, IsString, Matches, MaxLength } from 'class-validator';
@@
   `@ApiProperty`()
   `@IsString`()
   `@IsNotEmpty`()
+  `@MaxLength`(128)
   name: string;
@@
   `@ApiProperty`()
   `@IsString`()
   `@IsNotEmpty`()
+  `@MaxLength`(64)
+  `@Matches`(/^[a-z0-9_-]+$/)
   identifier: string;
@@
   `@ApiPropertyOptional`()
   `@IsString`()
   `@IsOptional`()
+  `@MaxLength`(1024)
   description?: string;

As per coding guidelines: apps/api changes should include proper input validation.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/dtos/create-agent-request.dto.ts` around lines 5 -
18, Add explicit length and format validation to the DTO properties: enforce a
sensible max length (e.g., 100) on name and identifier (use `@MaxLength` on both,
and optional `@MinLength` on name), and apply a URL-safe pattern for identifier
using `@Matches` (e.g., allow only lowercase letters, numbers,
hyphens/underscores) so identifiers are route-safe; also update the
corresponding `@ApiProperty/`@ApiPropertyOptional metadata to reflect these
constraints. Target the name and identifier fields in the CreateAgentRequest DTO
(the properties named "name" and "identifier") and add the class-validator
decorators and matching OpenAPI metadata descriptions accordingly.

apps/api/src/app/agents/mappers/agent-response.mapper.ts (1)

5-6: Add a blank line before each return statement.

This is a style-guideline mismatch only; behavior is fine.

As per coding guidelines: **/*.{ts,tsx,js,jsx}: “Include a blank line before every return statement.”

Also applies to: 18-19

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/mappers/agent-response.mapper.ts` around lines 5 - 6,
Add a blank line immediately before every return statement in this module to
satisfy the style guideline; specifically insert a blank line before the return
inside the toAgentResponse function and before the other return(s) in this file
(where other functions/methods produce AgentResponseDto or similar outputs) so
every return has a preceding empty line.

libs/dal/src/repositories/agent/agent.schema.ts (1)

32-33: Remove redundant index declarations.

These indexes duplicate field-level index: true declarations and add unnecessary index-management overhead.

♻️ Proposed cleanup

-agentSchema.index({ _organizationId: 1 });
-agentSchema.index({ _environmentId: 1 });
 agentSchema.index({ identifier: 1, _environmentId: 1 }, { unique: true });

As per coding guidelines: libs/dal/**: “Review with focus on data integrity and query performance.”

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@libs/dal/src/repositories/agent/agent.schema.ts` around lines 32 - 33, Remove
the redundant explicit index declarations agentSchema.index({ _organizationId: 1
}) and agentSchema.index({ _environmentId: 1 }) from agent.schema since those
fields already declare index: true at the field level; keep the field-level
index flags, delete the two agentSchema.index(...) lines, and quickly scan for
any other duplicate index declarations to avoid index-management overhead.

libs/dal/src/repositories/agent-integration/agent-integration.schema.ts (1)

41-42: Drop duplicate single-field indexes.

These duplicate field-level index: true declarations and are unnecessary.

♻️ Proposed cleanup

-agentIntegrationSchema.index({ _agentId: 1 });
-agentIntegrationSchema.index({ _environmentId: 1 });

As per coding guidelines: libs/dal/**: “Review with focus on data integrity and query performance.”

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@libs/dal/src/repositories/agent-integration/agent-integration.schema.ts`
around lines 41 - 42, The file currently adds duplicate single-field indexes via
agentIntegrationSchema.index({ _agentId: 1 }) and agentIntegrationSchema.index({
_environmentId: 1 }) while those fields already declare index: true; remove
these explicit index() calls so each field-level index is created only once
(delete the two agentIntegrationSchema.index(...) lines), and verify no required
compound indexes rely on those lines before committing.

apps/api/src/app/agents/usecases/add-agent-integration/add-agent-integration.usecase.ts (2)

30-37: Use array syntax for select for consistency.

Line 36 uses string syntax '_id' while other queries in this file use array syntax ['_id']. As per coding guidelines, array syntax should be the default for select.

Suggested fix

     const integration = await this.integrationRepository.findOne(
       {
         _id: command.integrationId,
         _environmentId: command.environmentId,
         _organizationId: command.organizationId,
       },
-      '_id'
+      ['_id']
     );

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@apps/api/src/app/agents/usecases/add-agent-integration/add-agent-integration.usecase.ts`
around lines 30 - 37, The select argument in the integrationRepository.findOne
call inside AddAgentIntegrationUseCase uses string syntax '_id' inconsistent
with the rest of the file; update that call (the findOne invocation in
add-agent-integration.usecase.ts) to use array syntax ['_id'] so it matches
other queries and coding guidelines (leave the query filter object unchanged).

---

63-65: Add blank line before return statement.

Per coding guidelines, include a blank line before every return statement.

Suggested fix

     const link = await this.agentIntegrationRepository.create({
       _agentId: agent._id,
       _integrationId: command.integrationId,
       _environmentId: command.environmentId,
       _organizationId: command.organizationId,
     });

+
     return toAgentIntegrationResponse(link);
   }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@apps/api/src/app/agents/usecases/add-agent-integration/add-agent-integration.usecase.ts`
around lines 63 - 65, Insert a blank line immediately before the final return in
add-agent-integration.usecase.ts so the `return
toAgentIntegrationResponse(link);` line is preceded by an empty line; locate the
return inside the add-agent-integration use case (referencing
`toAgentIntegrationResponse(link)` and the surrounding function/method in the
AddAgentIntegration use case) and add the blank line to satisfy the coding
guideline and linter formatting.

apps/api/src/app/agents/usecases/update-agent-integration/update-agent-integration.usecase.ts (3)

46-48: Add blank line before return statement.

Per coding guidelines, include a blank line before every return statement.

Suggested fix

     if (existingLink._integrationId === command.integrationId) {
+
       return toAgentIntegrationResponse(existingLink);
     }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@apps/api/src/app/agents/usecases/update-agent-integration/update-agent-integration.usecase.ts`
around lines 46 - 48, In update-agent-integration.usecase (the block comparing
existingLink._integrationId to command.integrationId) add a blank line
immediately before the return statement so the snippet reading "if
(existingLink._integrationId === command.integrationId) { return
toAgentIntegrationResponse(existingLink);" becomes separated by an empty line
prior to "return" to satisfy the coding guideline; locate this conditional in
the update-agent-integration.usecase function and insert the blank line before
the return.

---

50-57: Use array syntax for select for consistency.

Line 56 uses string syntax '_id' while other queries use array syntax. Per coding guidelines, array syntax should be the default.

Suggested fix

     const integration = await this.integrationRepository.findOne(
       {
         _id: command.integrationId,
         _environmentId: command.environmentId,
         _organizationId: command.organizationId,
       },
-      '_id'
+      ['_id']
     );

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@apps/api/src/app/agents/usecases/update-agent-integration/update-agent-integration.usecase.ts`
around lines 50 - 57, In update-agent-integration.usecase.ts, the
integrationRepository.findOne call uses string select syntax '_id' which is
inconsistent with the project's array select convention; change the select
argument to use array syntax ['_id'] in the findOne invocation inside the
UpdateAgentIntegrationUsecase (the const integration = await
this.integrationRepository.findOne(...) call) so it matches other repository
queries.

---

101-103: Add blank line before return statement.

Per coding guidelines, include a blank line before every return statement.

Suggested fix

     if (!updated) {
       throw new NotFoundException(
         `Agent-integration link "${command.agentIntegrationId}" was not found after update.`
       );
     }

+
     return toAgentIntegrationResponse(updated);
   }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@apps/api/src/app/agents/usecases/update-agent-integration/update-agent-integration.usecase.ts`
around lines 101 - 103, Add a blank line immediately before the final return
statement "return toAgentIntegrationResponse(updated);" in the
update-agent-integration.usecase (the function that updates the agent
integration, e.g., the method wrapping this diff) so the return is separated by
an empty line per coding guidelines; locate the return line and insert one empty
line above it.

apps/api/src/app/agents/agents.controller.ts (3)

54-59: Consider adding @SdkGroupName decorator.

Per coding guidelines, SDK endpoints should be grouped with the @SdkGroupName decorator using . as the subresource separator. Consider adding @SdkGroupName('Agents') at the class level.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 54 - 59, Add the
`@SdkGroupName` decorator at the class level to group SDK endpoints; insert
`@SdkGroupName`('Agents') on the AgentsController class (near the existing
decorators such as `@RequireAuthentication`(), `@ApiTags`('Agents') or
`@Controller`('/agents')) so the controller (AgentsController) is correctly
grouped using '.' as the subresource separator.

---

81-95: Add blank line before return statement.

Per coding guidelines, include a blank line before every return statement.

Suggested fix

   createAgent(
     `@UserSession`() user: UserSessionData,
     `@Body`() body: CreateAgentRequestDto
   ): Promise<AgentResponseDto> {
+
     return this.createAgentUsecase.execute(

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 81 - 95, In the
createAgent method, add a single blank line immediately before the return
statement so the method body separates the setup (building
CreateAgentCommand.create with
userId/environmentId/organizationId/name/identifier/description) from the
return; update the createAgent function to include that blank line above the
existing return to satisfy the coding guideline.

---

141-163: List endpoint should support pagination.

Similar to listAgents, the listAgentIntegrations endpoint returns an unbounded array. Consider adding pagination support per coding guidelines.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 141 - 163, The
listAgentIntegrations endpoint returns an unbounded array; update it to support
pagination by adding query parameters (e.g., page/limit or offset/limit) to the
controller method signature, validate them via DTO or decorators, and pass them
into ListAgentIntegrationsCommand (extend the command to accept pagination
fields). Update the listAgentIntegrationsUsecase.execute call to forward
pagination, change the response type from AgentIntegrationResponseDto[] to a
paginated wrapper (e.g., PaginatedResponse<AgentIntegrationResponseDto>) and
adjust ApiResponse/@ApiOperation metadata accordingly, and ensure the use case
and repository layer honor limit/offset to return total/count and items.

apps/api/src/app/agents/usecases/list-agent-integrations/list-agent-integrations.usecase.ts (1)

38-40: Add blank line before return statement.

Per coding guidelines, include a blank line before every return statement.

Suggested fix

     { sort: { createdAt: -1 } }
     );

+
     return links.map(toAgentIntegrationResponse);
   }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In
`@apps/api/src/app/agents/usecases/list-agent-integrations/list-agent-integrations.usecase.ts`
around lines 38 - 40, Insert a blank line immediately above the return statement
that currently reads "return links.map(toAgentIntegrationResponse);" so there is
an empty line separating preceding logic from the return; locate the return in
the list-agent-integrations use case where toAgentIntegrationResponse is used
and add the blank line to conform to the coding guideline.

apps/api/src/app/agents/usecases/update-agent/update-agent.usecase.ts (2)

61-63: Add blank line before return statement.

Per coding guidelines, include a blank line before every return statement.

Suggested fix

     if (!updated) {
       throw new NotFoundException(`Agent with identifier "${command.identifier}" was not found.`);
     }

+
     return toAgentResponse(updated);
   }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/usecases/update-agent/update-agent.usecase.ts` around
lines 61 - 63, Add a single blank line immediately before the return statement
in the UpdateAgent use case function (the line returning
toAgentResponse(updated) in update-agent.usecase.ts); locate the return inside
the function (e.g., execute/handle method that produces `updated`) and insert
one empty line above `return toAgentResponse(updated);` to satisfy the coding
guideline.

---

17-24: Select only required fields for efficiency.

The query retrieves all fields ('*') but only existing._id is used in subsequent operations.

Suggested fix

     const existing = await this.agentRepository.findOne(
       {
         identifier: command.identifier,
         _environmentId: command.environmentId,
         _organizationId: command.organizationId,
       },
-      '*'
+      ['_id']
     );

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/usecases/update-agent/update-agent.usecase.ts` around
lines 17 - 24, The findOne call in update-agent.usecase.ts fetches all fields
('*') into existing but only existing._id is used later; change the query in
AgentRepository.findOne (the call that sets the existing variable) to request
just the _id field (e.g., ['_id']) to reduce data transfer and improve
performance, keeping the same filter ({ identifier: command.identifier,
_environmentId: command.environmentId, _organizationId: command.organizationId
}) and leaving subsequent logic in UpdateAgentUsecase unchanged.

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

apps/api/src/app/agents/dtos/cursor-pagination-query.dto.ts (2)

34-39: Add @IsEnum validation for orderDirection.

Without @IsEnum(DirectionEnum) validation, invalid direction values will pass through the validation pipeline. The DirectionEnum only supports ASC and DESC.

♻️ Proposed fix

+import { IsEnum, IsOptional, IsString } from 'class-validator';
-import { IsOptional, IsString } from 'class-validator';

   `@ApiPropertyOptional`({
     description: 'Direction of sorting',
     enum: DirectionEnum,
   })
+  `@IsEnum`(DirectionEnum)
   `@IsOptional`()
   orderDirection?: DirectionEnum;

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/dtos/cursor-pagination-query.dto.ts` around lines 34
- 39, Add enum validation for the orderDirection DTO property: update the
property declaration for orderDirection to include the class-validator decorator
`@IsEnum`(DirectionEnum) (alongside `@IsOptional`()) so only DirectionEnum values
(ASC/DESC) pass validation; target the orderDirection property and the
DirectionEnum symbol in cursor-pagination-query.dto.ts when applying the change.

---

7-23: Use @ApiPropertyOptional for optional properties.

Per coding guidelines, optional TypeScript properties (?) should use @ApiPropertyOptional rather than @ApiProperty with required: false.

♻️ Proposed fix

-  `@ApiProperty`({
+  `@ApiPropertyOptional`({
     description: 'Cursor for pagination indicating the starting point after which to fetch results.',
     type: String,
-    required: false,
   })
   `@IsString`()
   `@IsOptional`()
   after?: string;

-  `@ApiProperty`({
+  `@ApiPropertyOptional`({
     description: 'Cursor for pagination indicating the ending point before which to fetch results.',
     type: String,
-    required: false,
   })
   `@IsString`()
   `@IsOptional`()
   before?: string;

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/dtos/cursor-pagination-query.dto.ts` around lines 7 -
23, Replace the `@ApiProperty`(...) decorators on the optional DTO fields after
and before with `@ApiPropertyOptional` so the decorators match the TypeScript
optional properties; locate the after and before properties in
CursorPaginationQueryDto (or the DTO class in cursor-pagination-query.dto.ts)
and swap the `@ApiProperty` usage (and remove required: false) to
`@ApiPropertyOptional` for both fields.

apps/api/src/app/agents/agents.controller.ts (1)

128-152: Add @SdkGroupName decorator for integration subresource endpoints.

Per coding guidelines, SDK endpoints should be grouped with @SdkGroupName decorator using . as the subresource separator. The agent-integration endpoints should use Agents.Integrations grouping.

♻️ Proposed fix for addAgentIntegration

+import { SdkGroupName } from '@novu/application-generic';

   `@Post`('/:identifier/integrations')
+  `@SdkGroupName`('Agents.Integrations')
   `@ApiResponse`(AgentIntegrationResponseDto, 201)

Apply the same @SdkGroupName('Agents.Integrations') decorator to listAgentIntegrations, updateAgentIntegration, and removeAgentIntegration methods.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 128 - 152, Add the
missing `@SdkGroupName`('Agents.Integrations') decorator to the integration
subresource endpoints: annotate addAgentIntegration and the sibling methods
listAgentIntegrations, updateAgentIntegration, and removeAgentIntegration with
`@SdkGroupName`('Agents.Integrations') so they are grouped correctly in the SDK
(use the exact decorator name and string 'Agents.Integrations' as the
subresource separator).

apps/api/src/app/agents/agents.controller.ts (2)

77-313: ⚠️ Potential issue | 🟠 Major

Re-add @ExternalApiAccessible on the Agents API handlers.

These routes expose the public Agents surface, but none of the handlers are marked as externally accessible anymore. That will keep them out of the user-API-key / official-SDK path even though the controller is implementing public resource endpoints.

As per coding guidelines: Routes accessible via user API keys or the official SDK must use @ExternalApiAccessible decorator in NestJS controllers.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 77 - 313, The
controller methods that implement the public Agents surface (createAgent,
listAgents, addAgentIntegration, listAgentIntegrations, updateAgentIntegration,
removeAgentIntegration, getAgent, updateAgent, deleteAgent) are missing the
`@ExternalApiAccessible` decorator; re-add `@ExternalApiAccessible` above each of
those handler methods and ensure ExternalApiAccessible is imported where
decorators are declared so these routes are exposed to user API keys / the
official SDK.

---

100-126: ⚠️ Potential issue | 🟡 Minor

Add @SdkUsePagination to both list endpoints.

Both handlers already implement cursor-based pagination semantics, but the required SDK pagination decorator is still missing. That leaves the controller out of sync with the project’s list-endpoint contract.

As per coding guidelines: List endpoints must support pagination and use @SdkUsePagination decorator.

Also applies to: 154-185

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 100 - 126, Add the
missing `@SdkUsePagination` decorator to the controller's cursor-paginated list
handlers: annotate the listAgents method (the function named listAgents) and the
other list endpoint in the same controller with `@SdkUsePagination`, and
add/import SdkUsePagination at the top of the file so the decorator is
available; ensure the decorator is placed above the method decorators (e.g.,
above `@Get` and `@ApiOperation`) to align with the project's SDK pagination
contract.

apps/api/src/app/agents/agents.controller.ts (1)

128-152: Rename the integration subresource handlers to the standard controller verbs.

addAgentIntegration and removeAgentIntegration break the controller naming convention used elsewhere. createAgentIntegration / deleteAgentIntegration would match the required pattern and keep the API surface consistent.

As per coding guidelines: Controller method names follow the pattern: getEntityName, listEntityName, createEntityName, updateEntityName, deleteEntityName.

Also applies to: 215-242

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@apps/api/src/app/agents/agents.controller.ts` around lines 128 - 152, Rename
the nonstandard controller methods addAgentIntegration and
removeAgentIntegration to follow the controller naming convention
(createAgentIntegration and deleteAgentIntegration respectively); update the
method names on the controller class, any corresponding decorator-declared
handlers (e.g., `@Post`('/:identifier/integrations') and the delete route around
lines referenced), and adjust any internal references (e.g., tests, route
registration, or calls within this file) so the command creation still uses
AddAgentIntegrationCommand.create and RemoveAgentIntegrationCommand.create as
before but invoked from the newly named methods; ensure method signatures,
decorators, permission decorators (RequirePermissions), and return types remain
unchanged.