Generates evidence-based performance hypotheses by analyzing git history and code execution paths.
The Performance Theory Gatherer is a specialized Claude Code skill designed to systematically identify the root causes of performance regressions. By analyzing recent git history and mapping code paths, it generates up to five structured hypotheses backed by specific code evidence and confidence levels. This skill helps developers move from vague observations to actionable theories by ensuring every performance claim is grounded in actual repository changes and architectural patterns, adhering to strict requirements for evidence-based analysis while avoiding premature optimization suggestions.
주요 기능
01Generates up to five structured hypotheses per performance scenario
02Maps code execution paths using repository mapping and grep functionality
031,152 GitHub stars
04Assigns confidence levels (low, medium, high) to each proposed theory
05Analyzes recent git history to identify potential performance regressions
06Strictly adheres to performance contracts defined in project documentation
사용 사례
01Developing an evidence-based performance profile for legacy codebases
02Investigating sudden latency spikes following recent code merges
03Validating performance theories before committing resources to optimization tasks
