efficiency-auditlisted

# Efficiency Audit Find the **1–3 most serious efficiency problems** in a codebase, prove they matter with a **concrete cost analysis**, and give the fix. These are the defects that dominate latency, cloud bill, or memory — the O(n²) inner loop, the N+1 query behind a list endpoint, the unbounded result set loaded into memory, the missing cache on an expensive repeated call. The whole point is **signal over volume**. A ruthless efficiency auditor doesn't flag every micro-optimization; it finds the one path that grows with your data or traffic and will fall over in production, quantifies the cost, and stops. Stopping early is a feature: the caller asked for the top few, and every extra finding past the target dilutes attention and burns the budget you were told to cap. **Cost must scale** — a slow path that runs once at startup on fixed-size input is not a finding; one whose cost grows with users, rows, requests, or time is. ## How it works This skill runs a **capped multi-agent workflow** (the `Workflow` tool) so recon, hunting, and verification happen in parallel across cheap **Sonnet** agents instead of serially in your own context. You stay the orchestrator: you launch the workflow, read its structured result, and write the report. The user invoking this skill is your explicit authorization to call `Workflow`. Three phases, hard-capped at roughly ten agents total: 1. **Recon** (1 agent) — map the *hot paths*: request handlers and their query patterns, loops over c