agents behavior

Exclude hanging Ziggurat benchmarks (infinite loop in spire)

Show confidence intervals and significance in comparison table

Fix WORK_DIR not being passed to Python comparison script

Lower default JMH settings to avoid multi-hour runs

Fix unbound variable error when no extra args are passed

Add script to benchmark goron's effect on spire

claude -> agents

Resolve goron automatically in sbt plugin via ivy dependency

Add sbt-goron plugin extending sbt-assembly

Add TODOs: benchmark ScalaInlineInfo impact, link-time specialization

Switch runtime benchmarks to AverageTime mode with proper warmup

Add AWFY runtime benchmarks for measuring goron's impact on compiled code

Add TODO: learn from GraalVM JIT optimizations and research papers

Fix Java 8 compatibility: replace Java 9+ and 11+ APIs