Repair Graceful Shutdown Dependency Bugs

The problem

Graceful shutdowns often fail due to race conditions, leaked background tasks, or incorrect dependency closing order. Developers struggle to debug these intermittent failures that only surface during process termination or high-concurrency drains.

What it does

• Identifies and fixes stop-accepting and drain phase inconsistencies in repository code.
• Ensures background task cancellation and in-flight request ownership are handled correctly.
• Enforces proper dependency close ordering to prevent null pointer exceptions or leaked resources.
• Generates deterministic local fixtures and regression tests to verify shutdown logic without flaky sleeps.
• Preserves existing shutdown deadlines and durability guarantees while cleaning up resource leaks.

Frameworks & tools

Works with repository-level code across languages using Read, Edit, and Bash tools. Focuses on systems handling process signals, concurrency, and async task management.

Why this beats prompting it yourself

Mapping complex dependency lifecycles across multiple files is error-prone for manual prompting. This skill systematically reconstructs state ownership and concurrent timelines to find the root cause of shutdown bugs rather than just wrapping calls in broad try-catch blocks.

Use cases

• Fixing hanging processes that fail to exit after receiving SIGTERM.
• Resolving database connection errors caused by closing pools before draining workers.
• Ensuring cleanup logic runs reliably during forced exits and grace period timeouts.
• Adding observability to track shutdown progress and identify stalled components.

Known limitations

Requires access to local repository files. Cannot interact with live production services or mutate deployed infrastructure state.