What problem does it solve? Manually auditing a large codebase for security holes, performance bottlenecks, accessibility gaps, outdated dependencies, and code quality issues takes days of engineering time and is rarely exhaustive. This Skill runs an autonomous multi-agent audit over a long runtime window and delivers a single severity-rated report. ## Core Features & Use Cases - Parallel Specialist Agents: Deploys five audit agents (Security, Performance, Accessibility, Dependency, Code Quality) simultaneously, each writing structured findings to its own report file. - Severity-Rated Findings: Every issue is rated Critical/High/Medium/Low using a shared rubric, then deduplicated and compiled into an executive summary with a top-10 priority list and sprint plan. - Scales to Large Repos: Handles codebases from under 10K lines to 500K+ lines by spawning sub-agents for file-batch parallelization. - Use Case: Point it at a legacy monorepo before a major refactor; wake up to overnight-audit-report.md containing every SQL injection risk, N+1 query, WCAG violation, and vulnerable dependency ranked by severity. ## Quick Start Run an overnight audit of this repository covering security, performance, accessibility, dependencies, and code quality, and write the compiled report to overnight-audit-report.md.
