ln-624-code-quality-auditor_skill

Overview

This skill performs a Level-3 code quality audit focused on complexity, method signatures, algorithmic efficiency, and constants management. It scans a specified path in the codebase and returns structured findings with severity, location, estimated effort, and concrete recommendations. It also computes a Code Quality compliance score and writes a single atomic markdown report to the configured output directory.

How this skill works

The worker parses the provided contextStore to determine scan_path, domain_mode, current_domain, and output_dir. It runs nine checks (cyclomatic complexity, deep nesting, long methods, god classes, too many parameters, quadratic algorithms, N+1 queries, magic numbers/constants, and method signature quality) scoped to scan_path. Findings are collected, tagged with domain when domain-aware, used to compute a penalty-based score, and emitted as a single markdown report file; a short summary is returned to the coordinator.

When to use it

• Integrating into an automated audit pipeline before release to catch maintainability and performance risks
• Running focused, domain-aware reviews when a microservice or feature folder is being evaluated
• Validating refactors to ensure complexity and signature quality improved
• Spot-checking for N+1 queries or O(n²) hotspots in API handlers and batch jobs
• Enforcing constants consolidation and reducing magic numbers before public API changes

Best practices

• Run in domain-aware mode to limit noise and focus on the relevant code slice
• Combine tool outputs (ESLint/radon/gocyclo) with simple pattern scans for better coverage
• Exclude tests, examples, and config files from magic-number detection
• Treat findings as guidance—do not auto-fix; include estimated effort for planning
• Prioritize CRITICAL/HIGH issues in hot paths (API endpoints, schedulers) for remediation

Example use cases

• Pre-merge audit of a service directory to catch N+1 queries and O(n²) loops introduced in a feature branch
• Monthly code-quality gate that produces a score and a single markdown report in the project audit folder
• Targeted review of a legacy module to identify god classes and long methods for refactor planning
• Verifying parameter and signature improvements after an API redesign
• Scanning a domain folder to consolidate duplicated constants into a single constants file