skill-auto-evolver

To identify performance bottlenecks, you can use execution tracing to collect telemetry like durations and error patterns. The analyzer then computes P50/P95/P99 latencies to surface resource-intensive execution paths for optimization.

The best way to run A/B experiments for service optimization is to generate data-driven suggestions, create optimized versions of your skill, and validate improvements using a managed experiment workflow with regression test baselines.

You can set up non-intrusive execution tracing using decorators that capture start/end timestamps and durations. This telemetry is stored locally in a lightweight SQLite database for subsequent latency percentile analysis.

Yes, you can manage regression testing baselines using a Python-based CLI tool backed by SQLite. It stores execution metrics and error statistics to validate skill improvements and detect performance regressions.

No, this performance monitoring tool does not require external databases. It stores execution metrics, percentile latencies, and optimization plans locally in SQLite, requiring only the Python click dependency for CLI tooling.