state-processor skill
Creates state processors to handle persistence and side effects for write operations like POST, PUT, PATCH, and DELETE.
- GitHub stars
- 29
- Files
- 1
- Last commit
- Oct 22, 2026
- License
- —
- Allowed tools
- —
What does the state-processor skill do?
state-processor creates state processors to handle persistence and side effects for write operations like POST, PUT, PATCH, and DELETE.
Summary generated from the skill's SKILL.md.
About the state-processor skill
Managing write operations and side effects in API Platform is what this skill enables. It allows developers to create state processors that handle the Command side of CQRS-style operations, specifically for POST, PUT, PATCH, and DELETE requests. When an agent needs to implement custom persistence logic, such as hashing passwords, creating related entities, or triggering email notifications, it can use this skill to define how data is processed during the write phase.
Developers can use these processors to decorate built-in services like the Doctrine ORM persist processor or the MongoDB ODM remove processor. This approach allows for adding logic before or after the standard persistence cycle. The skill also supports advanced use cases like implementing soft-delete functionality or returning Symfony Response objects for file downloads by setting the write flag to true on GET operations. It works across different frameworks, including Symfony and Laravel Eloquent, by utilizing the ProcessorInterface.
What an agent can do with it
- 1Implement custom side effectsAdd logic to send an email notification immediately after an order is successfully persisted.
- 2Handle soft-delete logicCreate a reusable processor that marks a resource as deleted instead of removing it from the database.
- 3Generate file downloadsConfigure a GET operation to trigger a processor that returns a binary file response.
- 4Automate related entity creationEnsure that child resources like mailboxes are automatically created whenever a new account is registered.
Written from the skill's SKILL.md. Every use case is checked against a quote from it.
How to install the state-processor skill
Each command copies the skill folder into the client's skills directory. Restart the client or start a new session to load it.Install the state-processor skill in Plugin
/plugin marketplace add api-platform/skillset /plugin install api-platform@api-platform-skillset
Install the state-processor skill in Claude Code
~/.claude/skills/state-processor/SKILL.md
$ npx degit api-platform/skillset/skills/state-processor ~/.claude/skills/state-processorInstall the state-processor skill in Claude Code (project)
.claude/skills/state-processor/SKILL.md
$ npx degit api-platform/skillset/skills/state-processor .claude/skills/state-processorInstall the state-processor skill in Codex
~/.agents/skills/state-processor/SKILL.md
$ npx degit api-platform/skillset/skills/state-processor ~/.agents/skills/state-processorInstall the state-processor skill in Gemini CLI
~/.gemini/skills/state-processor/SKILL.md
$ npx degit api-platform/skillset/skills/state-processor ~/.gemini/skills/state-processorInstall the state-processor skill in Claude.ai
npx degit api-platform/skillset/skills/state-processor state-processor && zip -r state-processor.zip state-processor # claude.ai → Customize → Skills → + → Create skill → Upload a skill → state-processor.zip
SKILL.md
View on GitHubCreating State Processors
State Processors handle persistence and side effects for write operations (Post, Patch, Put, Delete) — the write phase of an operation, the Command side of API Platform's CQRS-style split (a state-provider handles the Query/read side).
A processor runs when the operation's write flag is true. You normally leave it
unset and API Platform resolves it from the request: write defaults to "the HTTP
method is not safe", and read (the provider) to "the operation has URI
variables, or the method is safe".
| Operation | processor runs (write) |
provider runs (read) |
|---|---|---|
Get (item) / GetCollection |
no | yes |
Post (collection) |
yes | no (no URI vars, unsafe) |
Put, Patch, Delete (item) |
yes | yes |
A Get defaults to write: false, so its processor never fires until you set
write: true (file downloads, report generation — see Returning HTTP Responses
below). See operations → Read & write phases for the full read → deserialize → validate → write → serialize lifecycle.
Basic Structure
<?php
namespace App\State;
use ApiPlatform\Metadata\Operation;
use ApiPlatform\State\ProcessorInterface;
/**
* @implements ProcessorInterface<YourResource, YourResource>
*/
final class YourProcessor implements ProcessorInterface
{
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): mixed
{
// Business logic + persistence
return $data;
}
}
Decorating Built-in Processors
Wrap the default processor to add side effects before/after persistence:
use Symfony\Component\DependencyInjection\Attribute\Autowire;
final class OrderCreateProcessor implements ProcessorInterface
{
public function __construct(
#[Autowire(service: 'api_platform.doctrine.orm.state.persist_processor')]
private ProcessorInterface $persistProcessor,
private MailerInterface $mailer,
) {}
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): mixed
{
// Pre-persistence: set relationships, hash passwords, etc.
$data->setCreatedBy($this->getUser());
$result = $this->persistProcessor->process($data, $operation, $uriVariables, $context);
// Post-persistence: send notifications, trigger events
$this->mailer->send(new OrderConfirmation($result));
return $result;
}
}
Doctrine service names
| Persistence | Persist Processor | Remove Processor |
|---|---|---|
| ORM | api_platform.doctrine.orm.state.persist_processor |
api_platform.doctrine.orm.state.remove_processor |
| MongoDB ODM | api_platform.doctrine_mongodb.odm.state.persist_processor |
api_platform.doctrine_mongodb.odm.state.remove_processor |
Laravel (Eloquent):
ProcessorInterfaceis identical and resources reference a processor by class-string (processor: YourProcessor::class). The built-in Eloquent processors are bound in the container by class name —ApiPlatform\Laravel\Eloquent\State\PersistProcessor(handles create/update, including BelongsTo/HasMany relations andstandard_put) and…\State\RemoveProcessor. To decorate, type-hint that concrete class in your constructor instead of using a#[Autowire(service: 'api_platform.doctrine.*')]attribute. Side-effect patterns (mailers, events) work the same via Laravel's container. Returning a SymfonyResponsefrom a processor is also supported.
Creating Related Entities in a Processor
Create child resources alongside the main resource:
final class AccountCreateProcessor implements ProcessorInterface
{
public function __construct(
#[Autowire(service: 'api_platform.doctrine_mongodb.odm.state.persist_processor')]
private ProcessorInterface $persistProcessor,
) {}
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): mixed
{
$this->hashPassword($data);
// Create related entities before persisting
foreach (['INBOX', 'Sent', 'Trash', 'Drafts'] as $name) {
$mailbox = new Mailbox();
$mailbox->path = $name;
$data->addMailbox($mailbox); // cascade: ['persist']
}
// Single persist call saves parent + children
return $this->persistProcessor->process($data, $operation, $uriVariables, $context);
}
}
Soft-Delete Processor
Reusable processor for soft-deleting resources:
/**
* @implements ProcessorInterface<Account|Domain, Account|Domain>
*/
readonly class SoftDeleteProcessor implements ProcessorInterface
{
public function __construct(private DocumentManager $dm) {}
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): mixed
{
$data->isDeleted = true;
$this->dm->persist($data);
$this->dm->flush();
return $data;
}
}
Assign to Delete operations:
new Delete(processor: SoftDeleteProcessor::class)
Combine with a collection extension that filters isDeleted = false to hide soft-deleted records from queries (see securing-collections skill).
Returning HTTP Responses (File Downloads)
Processors can return a Symfony Response for binary content:
use Symfony\Component\HttpFoundation\HeaderUtils;
use Symfony\Component\HttpFoundation\Response;
/**
* @implements ProcessorInterface<Message, Response>
*/
final class DownloadProcessor implements ProcessorInterface
{
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): Response
{
$content = $this->buildContent($data);
$response = new Response($content);
$response->headers->set('Content-Type', 'application/octet-stream');
$response->headers->set('Content-Disposition',
HeaderUtils::makeDisposition(HeaderUtils::DISPOSITION_ATTACHMENT, $data->getId() . '.bin')
);
return $response;
}
}
Use with a Get operation that has write: true:
new Get(
uriTemplate: '/orders/{id}/download',
write: true,
processor: DownloadProcessor::class,
name: '_api_order_download',
)
Handling Both Persist and Delete
When a single processor handles Create/Update AND Delete, inject both services:
public function __construct(
#[Autowire(service: 'api_platform.doctrine.orm.state.persist_processor')]
private ProcessorInterface $persistProcessor,
#[Autowire(service: 'api_platform.doctrine.orm.state.remove_processor')]
private ProcessorInterface $removeProcessor,
) {}
public function process(mixed $data, Operation $operation, array $uriVariables = [], array $context = []): mixed
{
if ($operation instanceof DeleteOperationInterface) {
return $this->removeProcessor->process($data, $operation, $uriVariables, $context);
}
return $this->persistProcessor->process($data, $operation, $uriVariables, $context);
}
Processor Parameters
- $data: The deserialized/validated object
- $operation: Operation metadata (Post, Patch, Delete, etc.)
- $uriVariables: URI path variables
- $context: Includes
previous_data(before modification),request,resource_class
Best Practices
- Always return data from processors — omitting breaks serialization
- Decorate built-in processors rather than replacing them
- Keep processors focused: one processor per operation when logic differs significantly
- Use
$context['previous_data']in Patch/Put to detect what changed - Throw
HttpExceptionsubclasses for domain errors (BadRequestHttpException,UnprocessableEntityHttpException)
Files in this skill
1state-processor skill questions
How do I install the state-processor skill in Claude Code?
Run npx degit api-platform/skillset/skills/state-processor ~/.claude/skills/state-processor. Claude Code loads skills from ~/.claude/skills (all projects) or .claude/skills (one project) and uses a skill when its description matches the task.
Which clients can use the state-processor skill?
Any client that reads the Agent Skills format: Claude Code, Codex, Gemini CLI and claude.ai (upload as a ZIP). Scripts in the skill run only where the client allows code execution.
When does a state processor run?
A processor runs when the operation's 'write' flag is set to true, typically during POST, PUT, PATCH, or DELETE operations.
Can a processor be used for GET operations?
Yes, but you must explicitly set 'write: true' on the GET operation, which is useful for tasks like file downloads or report generation.
How do I add side effects like sending emails during a write operation?
You can decorate the built-in persistence processor to execute logic before or after the data is saved.
What is the difference between a provider and a processor in API Platform?
A state-provider handles the Query (read) side, while a state-processor handles the Command (write) side of the CQRS-style split.