Results for “adversary-simulation”

Validate skill outputs against assertion manifests to detect regressions, enforce quality gates, and bootstrap starter manifests from historical runs.

Execute the full AgentHub competition lifecycle in a single command: initialize, capture baseline, spawn agents, evaluate results, and merge the winner.