hammer-benchmark-regression

Performance regression detection compares current benchmark metrics against established baselines to identify degradation. This Skill runs a complete benchmark suite on SDL3 HammerEngine, extracts AI, collision, pathfinding, and rendering metrics, then flags regressions by severity level with likely causes and generates a detailed Markdown report for actionable insights.

The Skill executes all six critical HammerEngine benchmarks: AI Scaling, collision detection, pathfinding, event throughput, particle rendering, and UI stress testing. It extracts both synthetic and integrated metrics from each, with mandatory verification of AI and pathfinding results to ensure comprehensive performance coverage.

Yes. This is a core use case: run the full regression check on your feature branch before merge to catch performance degradation in AI, collision, or pathfinding systems early. It prevents bottlenecks from reaching the main branch and provides detailed reports identifying regression sources and severity.

You need Python 3 installed and access to the SDL3 HammerEngine codebase. The Skill handles full benchmark suite execution and metric extraction automatically, so no additional components or manual benchmark configuration is required beyond having Python 3 available.

Regressions are flagged with three severity levels—Critical, Warning, and Minor—based on metric deviations from baseline. The report identifies whether the source is AI infrastructure, integration, rendering, or other subsystems, enabling targeted optimization and preventing silent performance degradation.

Automated detection catches unintended performance degradation consistently before it reaches production. Manual testing is slow and error-prone; automation provides rigorous, repeatable regression detection with comprehensive baseline comparison across all six engine benchmarks on every run.