Last released Aug 1, 2026
AI that double-checks itself before it speaks: a skeptic harness that attacks LLM claims against source documents.
Supported by
