Spot performance drift in your benchmark runs.
What it does
Reports performance and cost drift across saved benchmark runs, highlighting win-rate and latency regressions.
When to use it
Use before a release, after expanding the benchmark corpus, or after upgrading agent-booster.
How to use it
Give it a project with saved benchmark runs; it reports first-to-last changes, a per-run series, and regression flags.
What you provide
- An existing project
Latest review
What you need · 1
Requires Node.js to run the trend script from the project root; no version is specified.
Access · 0
No additional actions listed.