Trust-aware memory for AI agents: evidence-backed recall, deterministic verdicts, self-inspection, and a tamper-evident local ledger.
Arakiss / nahuali
