perf-review

Performance Review

Analyse the specified files for performance issues and provide actionable improvement suggestions.

Files and instructions: $ARGUMENTS

What to review

• Algorithmic complexity: inefficient loops, nested iterations that could be flattened, O(n^2) or worse patterns that could be reduced, unnecessary repeated computations
• Batching: operations that could be grouped (database queries, API calls, file I/O) instead of executed one at a time in a loop
• Caching: repeated expensive computations or lookups that would benefit from memoisation or caching, missing cache invalidation
• Data structures: use of the wrong data structure for the access pattern (e.g. linear search in a list where a set or map would be O(1), arrays where linked lists are better or vice versa)
• Memory: unnecessary allocations, large objects held longer than needed, string concatenation in loops, unbounded collections, missing pagination
• Concurrency: sequential operations that could run in parallel, blocking calls on the main thread, missing async/await, thread-safety issues that force unnecessary serialisation
• I/O efficiency: chatty network calls, missing connection pooling, unbuffered reads/writes, missing streaming for large payloads
• Lazy vs eager evaluation: eagerly loading or computing data that may never be used, missing short-circuit evaluation, loading entire datasets when only a subset is needed
• Hot paths: performance-critical code paths where micro-optimisations matter — unnecessary object creation, redundant type conversions, avoidable regex compilation
• Language-specific antipatterns: idioms that are idiomatic but slow in the specific language (e.g. reflection in Java, dynamic dispatch where static would suffice, unintentional boxing)

How to interpret arguments

The arguments are free-form and flexible. They may contain:

• File references of any type and in any format: @file.ts, file.py, main.go, utils.go, script.sh handler.rb
• Additional natural language instructions alongside file references, such as:
  • "and also review the files imported by the specified file"
  • "focus only on database query performance"
  • "this handles 10k requests per second"
  • "we're seeing high memory usage in production"