What problem does it solve? Finding performance bottlenecks in a codebase is slow and error-prone when done manually, and reading code first creates anchoring bias that hides better solutions. This Skill runs a deep performance-only audit that systematically hunts for anti-patterns across 13 domains without being misled by existing implementations. ## Core Features & Use Cases - Parallel Multi-Agent Audit: Spawns 13 specialist agents simultaneously, each covering a performance domain such as database queries, memory leaks, algorithmic complexity, concurrency, bundle size, caching, and rendering. - Pattern-Based Detection: Agents use Grep/Glob pattern matching guided by reference files instead of reading full source code, avoiding anchoring bias on existing implementations. - Prioritized Consolidated Report: Deduplicates findings, sorts by severity (CRITICAL to LOW), groups by file, and produces an executive summary plus an ordered improvement plan with estimated impact. - Use Case: Ask to "find performance issues in my app" and receive a report flagging an N+1 query in your ORM loop, sequential awaits that should be Promise.all, and a missing database index — each with a concrete fix and expected impact. ## Quick Start Run a performance optimization audit on this project and report all bottlenecks with prioritized fixes.
