ProofAgent Harness (arXiv:2605.24134) by Fouad Bousetouane is the open source framework for adversarial AI agent evaluation. It runs a full pipeline of planning, adversarial conducting, multi juror scoring, debate consensus, and signed reporting, and returns a readiness verdict you can ship to security and audit teams.
Most evaluation scores a single answer with one scoring model. Production agents fail differently: three turns in, under pressure, through domain specific failure modes. The Harness selects domain relevant traps, sustains adversarial pressure across a full conversation, scores the whole trajectory with a panel of jurors, and produces evidence linked findings rather than one opaque number.
Capability is no longer the bottleneck. The risk that keeps agents out of production is behavior under pressure: a support agent that leaks data when pushed, a finance agent that breaks policy when flattered, a tool call that fires when it should refuse. Run the Harness as a gate in CI and CD, block deployment on a failed critical metric, and ship with a signed readiness report your auditors accept.
Across four domains and 25 adversarial turns, agents on GPT 5.5 sat at 7.5 out of 10 on both safety and manipulation resistance while agents on Claude Opus 4.7 reached 9.5 on safety. Failures are metric-specific, and one domain moved 2.4 points depending on which Harness regime graded it. Every result is reproducible from the open source code under Apache 2.0. The companion paradigm paper, Human-on-the-Bridge (arXiv:2606.16871), formalizes the approach.
Plain language explainer on Medium: ProofAgent Harness: The Open Source Harness for Complete AI Agent Evaluation.