Background Job Reliability Auditor

The problem

Background jobs and queue consumers often fail silently or cause duplicate side effects like double-billing or multiple email sends. Standard code reviews usually miss the complex edge cases of distributed systems, such as visibility timeouts, race conditions, and partial failures.

What it does

• Analyzes worker implementations for idempotency and duplicate-processing risks.
• Audits retry policies, backoff strategies, and dead-letter queue handling.
• Maps execution flows to identify points of potential data loss or state corruption.
• Evaluates concurrency controls, locking mechanisms, and database transaction boundaries.
• Generates a 100-point reliability score and a prioritized remediation plan.

Why this beats prompting it yourself

General-purpose LLMs lack the specific rubric for distributed systems reliability. This skill follows a rigorous 20-step audit workflow that forces checks for often-forgotten failure modes like lease expiry, clock drift in schedulers, and response ambiguity in external APIs.

Use cases

• Reviewing an asynchronous billing or payment processing job before launch.
• Auditing a high-volume queue consumer for backlog and starvation risks.
• Validating the safety of a manual job replay or recovery procedure.
• Defining the specific test cases needed to prove a worker is production-ready.

Known limitations

Cannot execute workers or connect to live brokers. It performs static analysis and requires you to provide the implementation code or configuration.