Performance Reviewer

Overview

Performance Reviewer is a specialized AI agent designed to rigorously analyze codebases for potential performance bottlenecks. Its primary mission is to identify hotspots, assess algorithmic complexity (e.g., identifying O(n^2) patterns), and review resource usage like memory allocation and I/O latency.

This tool focuses purely on efficiency and scalability, deliberately excluding style checks or general logic validation. It operates under the principle that performance issues often compound silently until they become critical production incidents.

Capabilities

• Hot Path Identification: Pinpoints sections of code executed frequently or dealing with large datasets.
• Algorithmic Complexity Analysis: Detects inefficient patterns like nested loops, repeated searches within hot paths, and suboptimal sorting implementations.
• Resource Pattern Review: Examines memory usage (e.g., allocations in tight loops, excessive string concatenation) and I/O blocking patterns.
• Data-Driven Recommendations: Provides quantifiable findings rather than vague warnings, recommending profiling steps before suggesting specific optimizations.

Example Use Cases

• Scaling Assessment: Analyzing a function that processes user records to determine if an O(n^2) complexity will fail when the dataset grows from 100 to 10,000 entries.
• Loop Optimization: Reviewing a data processing loop suspected of high memory churn due to repeated object creation or string concatenation.
• Bottleneck Pinpointing: Given a service endpoint that times out under load, this agent can help narrow down whether the bottleneck is CPU-bound (algorithmic) or I/O-bound (latency).

Note: The agent prioritizes evidence-dense output and will guide you toward profiling first, unless an algorithmic flaw is immediately obvious.