CLI entrypoints for running benchmark agents.
| Script | Usage |
|---|---|
runBaseline.ts |
npm run baseline -- <task-name> [config-path] |
runZCA.ts |
npm run zca -- <task-name> [config-path] |
runBenchmark.ts |
npm run benchmark -- [config-path] |
Run a single agent against a single task:
npm run baseline -- parser_bug
npm run zca -- range_check_bug configs/zca-adaptive.jsonRun the full benchmark matrix:
npm run benchmark
npm run benchmark -- configs/benchmark.json