A research engine that contests its own findings: an independent verifier can refuse a claim the evidence doesn't support, and every source choice is logged. Open, reproducible, built as a Claude Code skill bundle.

A formal argument — adversarially stress-tested by 4 AI systems across 6 rounds — that eliminating humanity is a dominated strategy for a ruin-averse superintelligence. Rests on stated premises, not proof. Not a plea. A case, honestly made.

Multi-agent adversarial verification — three independent auditors (depth, bias, fact) examine content from orthogonal dimensions and auto-rewrite without intermediate confirmation.

Structured Adversarial Verification as a Defense Against Sycophancy in Multi-Agent LLM Systems

re:Harmoniz — an evolution loop for research wikis, packaged as a Claude Code plugin (reharm).