agent-steward

Quality gate + spend auditor for AI agent fleets: your rules as executable checks, measured spend from transcripts, and tier tuning that quality data can veto

GitHub

Official source

Start locally with the CLI guide