Can it identify memory leaks?

Yes, it audits resource management for issues like unclosed connections, event listeners that are never removed, and large objects held in closures.

How does the confidence scoring work?

Each finding is assigned a score from 0-100 based on the evidence; only findings with a score of 80 or above are reported to ensure high-quality, actionable feedback.

What does the Performance Review skill do?

It analyzes code to find bottlenecks like inefficient loops, database N+1 issues, and memory leaks using a structured, evidence-based approach that estimates real-world impact.

How does it handle database efficiency?

The skill identifies common I/O issues such as queries inside loops, missing pagination, and sequential calls that should be parallelized, providing before-and-after code examples for fixes.

What is the difference between a hot path and a cold path?

A hot path is code executed frequently where performance is critical for UX, while a cold path is rarely run (like migrations); the skill prioritizes optimizations for hot paths.

Performance Review Steward

Name: Performance Review Steward
Author: liminal-ai

byliminal-ai

0•

開発者ツール

Analyzes codebases to identify performance bottlenecks, inefficient algorithms, and resource management issues using evidence-based metrics.

The Performance Review skill provides a rigorous, three-phase framework for evaluating how code behaves under production load and scale. It moves beyond vague observations to provide concrete, evidence-based findings on algorithmic complexity, N+1 query patterns, and memory management. By assessing project context, comparing against established patterns, and categorizing issues into specific performance buckets, it ensures that optimizations are targeted, high-confidence, and actionable for high-traffic environments.

主な機能

01Memory leak and resource cleanup auditing for high-traffic applications

02Confidence scoring system to ensure only high-impact findings are reported

03Algorithmic complexity identification with concrete O(n) impact estimates

04Three-phase analysis covering context research, comparative patterns, and issue assessment

05Detection of N+1 query patterns and inefficient I/O operations

060 GitHub stars

ユースケース

01Pre-production performance auditing for scalable backend APIs

02Identifying the root cause of slow database queries and network overhead

03Reviewing Pull Requests specifically for scalability and resource leaks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add liminal-ai/code-steward performance-review

For use in Claude.ai and ChatGPT

主な機能

01Memory leak and resource cleanup auditing for high-traffic applications

02Confidence scoring system to ensure only high-impact findings are reported

03Algorithmic complexity identification with concrete O(n) impact estimates

04Three-phase analysis covering context research, comparative patterns, and issue assessment

05Detection of N+1 query patterns and inefficient I/O operations

060 GitHub stars

ユースケース

01Pre-production performance auditing for scalable backend APIs

02Identifying the root cause of slow database queries and network overhead

03Reviewing Pull Requests specifically for scalability and resource leaks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add liminal-ai/code-steward performance-review

For use in Claude.ai and ChatGPT