# Create Job ## When To Use Use this skill for asynchronous, slow, retryable, external, scheduled, or failure-tolerant work. ## Inputs Needed - Work to perform, payload, queue, timeout, retries, uniqueness, idempotency, failure handling, and tests. ## Workflow 1. Inspect existing queue conventions. 2. Create the job with Artisan. 3. Prefer queue attributes such as `#[Queue]`, `#[Connection]`, `#[Tries]`, `#[Backoff]`, `#[Timeout]`, `#[MaxExceptions]`, `#[Delay]`, `#[UniqueFor]`, `#[FailOnTimeout]`, `#[DeleteWhenMissingModels]`, and `#[WithoutRelations]` for simple job metadata. 4. Pass IDs or small serializable data instead of large mutable objects when practical. 5. Inject services into `handle()` instead of calling `app()` from the job body. 6. Set timeout, tries or retry strategy, backoff, and failure behavior. 7. Make retries idempotent where possible. 8. Dispatch after commit when the job depends on newly written data, or set queue connection `after_commit` behavior intentionally. 9. Add tests with queue fakes or synchronous execution as appropriate. ## Files That May Be Created - `app/Jobs/*` - Job tests. ## Files That May Be Modified - Actions, listeners, commands, schedules, queue config, and tests. ## Architecture Rules - Jobs run background work; Actions own primary business operations. - Jobs may call Actions or Services. - Jobs should not contain UI or HTTP response logic. - Job dependencies should be method-injected into `handle()` unless they must be serialized. ## Testing Requirements - Test dispatch, payload, retry-sensitive behavior, failure handling, and idempotency where relevant. ## Security Requirements - Re-authorize or validate assumptions for delayed work when permissions or state may change. - Avoid logging secrets. ## Review Checklist - Is async work necessary? - Is the job retry-safe? - Are timeout and retry settings explicit? ## Common Mistakes - Queueing non-idempotent work without safeguards. - Dispatching inside a transaction too early. - Storing user-provided secrets in job payloads.

Create Job

When To Use

Use this skill for asynchronous, slow, retryable, external, scheduled, or failure-tolerant work.

Inputs Needed

Work to perform, payload, queue, timeout, retries, uniqueness, idempotency, failure handling, and tests.

Workflow

Inspect existing queue conventions.
Create the job with Artisan.
Prefer queue attributes such as #[Queue], #[Connection], #[Tries], #[Backoff], #[Timeout], #[MaxExceptions], #[Delay], #[UniqueFor], #[FailOnTimeout], #[DeleteWhenMissingModels], and #[WithoutRelations] for simple job metadata.
Pass IDs or small serializable data instead of large mutable objects when practical.
Inject services into handle() instead of calling app() from the job body.
Set timeout, tries or retry strategy, backoff, and failure behavior.
Make retries idempotent where possible.
Dispatch after commit when the job depends on newly written data, or set queue connection after_commit behavior intentionally.
Add tests with queue fakes or synchronous execution as appropriate.

Files That May Be Created

app/Jobs/*
Job tests.

Files That May Be Modified

Actions, listeners, commands, schedules, queue config, and tests.

Architecture Rules

Jobs run background work; Actions own primary business operations.
Jobs may call Actions or Services.
Jobs should not contain UI or HTTP response logic.
Job dependencies should be method-injected into handle() unless they must be serialized.

Testing Requirements

Test dispatch, payload, retry-sensitive behavior, failure handling, and idempotency where relevant.

Security Requirements

Re-authorize or validate assumptions for delayed work when permissions or state may change.
Avoid logging secrets.

Review Checklist

Is async work necessary?
Is the job retry-safe?
Are timeout and retry settings explicit?

Common Mistakes

Queueing non-idempotent work without safeguards.
Dispatching inside a transaction too early.
Storing user-provided secrets in job payloads.