Can it suggest optimizations without data?

No, the skill is constrained to only cite evidence found in logs or code, ensuring recommendations are accurate and strictly data-driven.

Does it follow a specific performance protocol?

Yes, it is designed to follow 'docs/perf-requirements.md' as the canonical contract for all performance-related syntheses.

What format are the analysis results delivered in?

It provides a structured output including a summary, actionable recommendations, abandoned hypotheses, and suggested next steps.

How does it handle insufficient profiling data?

If the provided data is insufficient to reach a conclusion, the skill will explicitly state the deficiency and request a re-run of the analysis.

What data does the Performance Analyzer require?

It accepts baseline data, experiment results, profiling evidence, hypotheses tested, and breaking point results to generate insights.

Performance Analyzer

Name: Performance Analyzer
Author: agent-sh

byagent-sh

•

692

•

Analytics & Monitoring

Synthesizes performance investigation data into evidence-backed recommendations and actionable optimization decisions.

Performance Analyzer is a specialized skill for Claude Code designed to transform raw performance data, profiling evidence, and experimental results into structured, actionable reports. By analyzing baseline metrics against experimental hypotheses and breaking point results, it helps developers identify bottlenecks, document failed experiments, and determine the clear next steps for optimization. It strictly adheres to evidence-based reporting, ensuring that every recommendation is supported by logs or code analysis rather than speculation, following your project's specific performance requirements.

Key Features

01Identifies breaking points and baseline deviations in application performance

02Tracks and documents abandoned hypotheses to prevent redundant testing

03Enforces data-driven constraints to ensure recommendations are evidence-backed

04692 GitHub stars

05Generates actionable performance recommendations based on empirical data

06Synthesizes profiling evidence and experiment results into clear summaries

Use Cases

01Conducting post-mortem performance reviews after a system slowdown or crash

02Summarizing complex profiling data into human-readable insights and next steps

03Evaluating the success of optimization experiments against baseline metrics

Key Features

01Identifies breaking points and baseline deviations in application performance

02Tracks and documents abandoned hypotheses to prevent redundant testing

03Enforces data-driven constraints to ensure recommendations are evidence-backed

04692 GitHub stars

05Generates actionable performance recommendations based on empirical data

06Synthesizes profiling evidence and experiment results into clear summaries

Use Cases

01Conducting post-mortem performance reviews after a system slowdown or crash

02Summarizing complex profiling data into human-readable insights and next steps

03Evaluating the success of optimization experiments against baseline metrics