ln-629-lifecycle-auditor_skill

Overview

This skill audits application lifecycle and entry points to detect startup and shutdown defects that risk downtime. It inspects bootstrap order, graceful shutdown, resource cleanup, signal handling, and liveness/readiness probes, then produces a scored report and concise summary for the coordinator. The output is a single atomic markdown report written to the specified output_dir and a minimal summary returned to the coordinator.

How this skill works

The worker parses the coordinator-provided contextStore to learn the tech stack, deployment type, codebase root, and output_dir. It scans main/entry files and runtime wiring to validate initialization order and looks for signal handlers, shutdown logic, resource cleanup, and probe endpoints. Findings are collected with severity, location, estimated effort, and remediation recommendations, a compliance score is calculated, and a single markdown report is written to {output_dir}/629-lifecycle.md. A compact summary line is returned to the coordinator.

When to use it

• Before production deployments to validate lifecycle robustness
• During architecture reviews for services with long-running processes
• When adding or refactoring startup or shutdown logic
• For containerized apps to validate Kubernetes probes and signal handling
• As part of a delivery pipeline gate to prevent downtime-inducing regressions

Best practices

• Initialize in order: config → DB → routes/middleware → server
• Implement SIGTERM and SIGINT handlers that close DBs and stop workers
• Ensure all file handles, sockets, and threads are explicitly closed on shutdown
• Expose /live and /ready endpoints for Kubernetes; bind probes to simple checks
• Treat initialization order violations that use DB before initialization as HIGH severity

Example use cases

• Audit a Node.js microservice to ensure process.on('SIGTERM') exists and DB.close() is called
• Validate a Go server for signal.Notify usage and graceful termination of goroutines
• Confirm containerized services expose /live and /ready and that Kubernetes probes are configured
• Run as an L3 worker in a coordinator pipeline to gate deployments with a lifecycle score
• Detect missing resource cleanup in services that spawn worker threads or open files