Does it track failed experiments?

Yes, it specifically identifies and lists abandoned hypotheses or experiments that failed to meet performance goals.

How does it ensure the accuracy of recommendations?

The skill is strictly constrained to only cite evidence that exists in logs or code, preventing the AI from hallucinating performance improvements.

What data does the perf-analyzer skill require?

It synthesizes baseline data, experiment results, profiling evidence, hypotheses, and breaking point results into a clear report.

What happens if there is not enough performance data available?

If the data is insufficient for a clear synthesis, the skill will state so explicitly and request that the user re-run the performance tests.

Performance Analysis & Synthesis

Name: Performance Analysis & Synthesis
Author: avifenesh

byavifenesh

•

311

分析と監視

Synthesizes performance investigation data into evidence-backed recommendations and actionable next steps.

概要

The perf-analyzer skill empowers developers to transform raw performance data—including profiling evidence, experiment results, and baseline metrics—into structured, actionable reports. By strictly following performance requirements and citing only verified evidence from logs or code, it helps teams make informed decisions about system optimizations, identify failed hypotheses, and determine the most effective paths for future performance improvements within their codebase.

主な機能

Actionable recommendation generation
Evidence-backed performance synthesis
Hypothesis validation and abandonment tracking
Integration with performance requirement contracts
Automated next-step determination based on data
311 GitHub stars

ユースケース

Evaluating the success or failure of performance-related experiments against a baseline.
Analyzing profiling results after a load test to identify critical bottlenecks.
Consolidating disparate logging data into a clear executive summary for architectural decisions.

概要

主な機能

Actionable recommendation generation
Evidence-backed performance synthesis
Hypothesis validation and abandonment tracking
Integration with performance requirement contracts
Automated next-step determination based on data
311 GitHub stars

ユースケース

Evaluating the success or failure of performance-related experiments against a baseline.
Analyzing profiling results after a load test to identify critical bottlenecks.
Consolidating disparate logging data into a clear executive summary for architectural decisions.