Paths: File paths (references/, ../ln-*) are relative to this skill directory.
Runtime Lifecycle & Config Auditor (L3 Worker)
Type: L3 Worker
Specialized worker auditing runtime readiness, lifecycle, and startup configuration validation.
Purpose & Scope
Audit runtime lifecycle and config validation (Category 12: Medium Priority)
Check bootstrap, shutdown, signal handling, probes, resource cleanup, env/config sync, and fail-fast startup validation
Emit FIX_BOOTSTRAP, ADD_CONFIG_VALIDATION, or FIX_SHUTDOWN
Calculate compliance score (X/10)
Inputs
MANDATORY READ: Load references/audit_worker_core_contract.md and references/mcp_tool_preferences.md.
Tool policy: You may run as an isolated subagent where host AGENTS.md is not in scope, so default to hex-line MCP first for file reads, search, and edits. Load references/mcp_integration_patterns.md only when MCP behavior is unclear.
Receives contextStore with tech stack, deployment type, codebase root, output_dir.
Use hex-graph first when lifecycle tracing materially improves confidence. Use hex-line first for local code reads when available. If MCP is unavailable, unsupported, or not indexed, continue with built-in Read/Grep/Glob/Bash and state the fallback in the report.
Workflow
Detection policy: use two-layer detection (candidate scan, then context verification); load references/two_layer_detection.md only when the verification method is ambiguous.
Parse context + output_dir
Check lifecycle and config validation patterns (Layer 1: grep for SIGTERM, shutdown handlers, probes, env reads, settings validation)
Analyze context per candidate (Layer 2):
Bootstrap order: read main file -- trace actual init sequence, verify dependencies satisfied before use
Graceful shutdown: read signal handlers -- do they actually close all resources? Or just log and exit?
Resource cleanup: read shutdown handler -- are ALL opened resources (DB, Redis, queues) closed?
Probes: check deployment config (Dockerfile, k8s manifests) -- is this containerized?
Config validation: are required env/config values validated at startup, before serving traffic?
Collect confirmed findings
Calculate score
Write Report: Build full markdown report in memory per references/templates/audit_worker_report_template.md, write to {output_dir}/ln-629--global.md in single Write call
Return Summary: Return minimal summary
Audit Rules
1. Bootstrap Initialization Order
Detection:
Check main/index file for initialization sequence
Verify dependencies loaded before usage (DB before routes)
Severity:
HIGH: Incorrect order causes startup failures
Recommendation: Initialize in correct order: config -> DB -> routes -> server
Effort: M (refactor startup)
2. Graceful Shutdown
Detection:
Grep for SIGTERM, SIGINT handlers
Check process.on('SIGTERM') (Node.js)
Check signal.Notify (Go)
Severity:
HIGH: No shutdown handler (abrupt termination)
Recommendation: Add SIGTERM handler, close connections gracefully
Effort: M (add shutdown logic)
3. Resource Cleanup on Exit
Detection:
Check if DB connections closed on shutdown
Verify file handles released
Check worker threads stopped
Severity:
MEDIUM: Resource leaks on shutdown
Recommendation: Close all resources in shutdown handler
Effort: S-M (add cleanup calls)
4. Signal Handling
Detection:
Check handlers for SIGTERM, SIGINT, SIGHUP
Verify proper signal propagation to child processes
Severity:
MEDIUM: Missing signal handlers
Recommendation: Handle all standard signals
Effort: S (add signal handlers)
5. Liveness/Readiness Probes
Detection (for containerized apps):
Check for /live, /ready endpoints
Verify Kubernetes probe configuration
Severity:
MEDIUM: No probes (Kubernetes can't detect health)
Recommendation: Add /live (is running) and /ready (ready for traffic)
Effort: S (add endpoints)
6. Startup Config Validation
Detection:
Find env/config reads in startup path
Check for validation frameworks or explicit fail-fast checks
Compare required runtime vars from code with documented startup config when available
Severity:
HIGH: Required config can be missing while app still starts
MEDIUM: Defaults/desync can make local, CI, and deployed startup differ
Recommendation: Validate required config at boot and fail before accepting traffic
Write JSON summary per references/audit_summary_contract.md. In managed mode the caller passes both runId and summaryArtifactPath; in standalone mode the worker generates its own run-scoped artifact path per shared contract.
Write report to {output_dir}/ln-629--global.md with category: "Runtime Lifecycle & Config" and checks: bootstrap_order, graceful_shutdown, resource_cleanup, signal_handling, probes, startup_config_validation.
Return summary per references/audit_summary_contract.md.
When summaryArtifactPath is absent, write the standalone runtime summary under .hex-skills/runtime-artifacts/runs/{run_id}/evaluation-worker/{worker}--{identifier}.json and optionally echo the same summary in structured output.
