Resilient Check

Resilient Check

When to use

Use this skill for Laravel jobs, integrations, webhooks, scheduled tasks, listeners, notifications, or user-facing flows that must stay reliable when dependencies are slow, unavailable, or inconsistent. It is most useful when queues, Redis, Horizon, external APIs, or payment/provider callbacks are involved.

Input parameters

• The workflow or subsystem to assess.
• External services, queues, or webhooks involved.
• Existing retry, timeout, circuit-breaker, or deduplication behavior.
• Known incidents, flaky tests, or production failure modes.
• Relevant jobs, listeners, commands, controllers, events, notifications, or queue config files.

Procedure

1. Trace the Laravel execution path for the workflow. Inspect controllers, dispatched jobs, listeners, commands, notifications, service classes, and relevant config such as config/queue.php, config/services.php, and Horizon settings.
2. Identify where the workflow can fail or partially succeed: external HTTP calls, payment providers, webhooks, database writes, cache writes, file storage, notifications, and chained or batched jobs.
3. Review retry behavior in Laravel terms. Check tries, backoff, retryUntil, job timeouts, Horizon worker configuration, queue visibility/timeout alignment, and whether failures land in failed_jobs with useful context.
4. Review idempotency and duplicate-delivery safety. Look for webhook deduplication, cache locks, unique jobs, database uniqueness guarantees, firstOrCreate race conditions, unsafe side effects before persistence, and handlers that are not safe to replay.
5. Flag high-risk patterns such as synchronous third-party API calls inside controllers, jobs that both mutate state and call external APIs without compensation, unbounded retries, missing dead-letter handling, and scheduled tasks that can overlap without withoutOverlapping() or locks.
6. Return findings under High-risk failure modes, Code changes, and Operational safeguards. Prefer concrete Laravel fixes such as queue settings, locks, unique jobs, timeout changes, retry tuning, ThrottlesExceptions, or graceful fallback responses.
7. Keep the answer practical and prioritized. Focus on the smallest changes that reduce operational risk fastest.

Output expectations: return the highest-risk failure modes first, then concrete Laravel code and operational fixes, with the smallest high-impact next steps called out clearly.

Examples

Prompt 1:

Use resilient-check on our payment webhook processing flow. Focus on duplicate deliveries, idempotency, timeouts, retry storms, and safe recovery.

Prompt 2:

Review this Laravel job pipeline for resilience. We call two external APIs and workers sometimes retry the same job for hours. Prioritize the highest-risk failure modes first.

JSON:

{
  "skill": "resilient-check",
  "context": {
    "workflow": "payment_webhook_processing",
    "dependencies": ["stripe", "erp_api", "redis_queue"],
    "focus": ["idempotency", "timeouts", "retries", "fallbacks"],
    "files": ["app/Jobs", "app/Listeners", "config/queue.php"],
    "concerns": ["duplicate_delivery", "retry_storms", "timeout_alignment"]
  }
}

Code example:

final class SyncInvoiceJob implements ShouldQueue
{
    public int $tries = 5;
    public int $timeout = 30;
    public array $backoff = [60, 300, 900];
}

Smoke test

Use the skill on a Laravel queue-driven integration flow with synchronous API calls, naive retries, and no deduplication. Verify that the response references job-level settings like tries, backoff, and timeouts, flags idempotency gaps, and recommends concrete Laravel resilience fixes instead of general reliability advice.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: me-shaon
• Source: me-shaon/agent-skills
• License: MIT

Install and usage instructions live in the source repository linked above.