Runs the SIN-Code tool performance benchmarks and persists results to
benchmark_results/.
Provides a reproducible, scriptable way to measure latency / throughput of
each SIN-Code tool across the full suite (discover, execute, map,
grasp, scout, harvest, orchestrate, bundle, brain, sckg).
./benchmark.sh # run all benchmarks
./benchmark.sh discover # run a single toolbenchmark_results/<tool>_<timestamp>.json— machine-readable resultsbenchmark_results/<tool>_<timestamp>.md— human-readable summary
- Requires all SIN-Code CLIs on
$PATH - Cold-cache runs include tool import overhead; warm-cache runs are stable