# Refute-or-Promote: An Adversarial Stage-Gated Multi-Agent Review Methodology for High-Precision LLM-Assisted Defect Discovery

arXiv: https://arxiv.org/abs/2604.19049 · HTML: https://mlsanity.com/paper/2604.19049/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Web, search, and research agents, one of 16 directions in LLM agents and tool use.
- Builds on 28 of its 51 references that are in the map.
- First posted on arXiv 21 Apr 2026.

## Placed in

- LLM agents and tool use / [Web, search, and research agents](https://mlsanity.com/agents/web-search-and-research-agents/)

## Builds on (most cited first, 28 of 28)

- [Reflexion: Language Agents with Verbal Reinforcement Learning](https://mlsanity.com/paper/2303.11366/) (2023)
- [Self-Refine: Iterative Refinement with Self-Feedback](https://mlsanity.com/paper/2303.17651/) (2023)
- [Improving Factuality and Reasoning in Language Models through Multiagent Debate](https://mlsanity.com/paper/2305.14325/) (2023)
- [XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models](https://mlsanity.com/paper/2308.01263/) (2023)
- [Mixture-of-Agents Enhances Large Language Model Capabilities](https://mlsanity.com/paper/2406.04692/) (2024)
- [AI safety via debate](https://mlsanity.com/paper/1805.00899/) (2018)
- [Debating with More Persuasive LLMs Leads to More Truthful Answers](https://mlsanity.com/paper/2402.06782/) (2024)
- [OR-Bench: An Over-Refusal Benchmark for Large Language Models](https://mlsanity.com/paper/2405.20947/) (2024)
- [IRIS: LLM-Assisted Static Analysis for Detecting Security Vulnerabilities](https://mlsanity.com/paper/2405.17238/) (2024)
- [Talk Isn't Always Cheap: Understanding Failure Modes in Multi-Agent Debate](https://mlsanity.com/paper/2509.05396/) (2025)
- [Multi-Agent Verification: Scaling Test-Time Compute with Multiple Verifiers](https://mlsanity.com/paper/2502.20379/) (2025)
- [Rethinking Mixture-of-Agents: Is Mixing Different Large Language Models Beneficial?](https://mlsanity.com/paper/2502.00674/) (2025)
- [SWE-Debate: Competitive Multi-Agent Debate for Software Issue Resolution](https://mlsanity.com/paper/2507.23348/) (2025)
- [RepoAudit: An Autonomous LLM-Agent for Repository-Level Code Auditing](https://mlsanity.com/paper/2501.18160/) (2025)
- [Correlated Errors in Large Language Models](https://mlsanity.com/paper/2506.07962/) (2025)
- [Let the Trial Begin: A Mock-Court Approach to Vulnerability Detection using LLM-Based Agents](https://mlsanity.com/paper/2505.10961/) (2025)
- [Beyond Consensus: Mitigating the Agreeableness Bias in LLM Judge Evaluations](https://mlsanity.com/paper/2510.11822/) (2025)
- [Debate, Deliberate, Decide (D3): A Cost-Aware Adversarial Framework for Reliable and Interpretable LLM Evaluation](https://mlsanity.com/paper/2410.04663/) (2024)
- [Consensus is Not Verification: Why Crowd Wisdom Strategies Fail for LLM Truthfulness](https://mlsanity.com/paper/2603.06612/) (2026)
- [VulAgent: Hypothesis-Validation based Multi-Agent Vulnerability Detection](https://mlsanity.com/paper/2509.11523/) (2025)
- [InfCode: Adversarial Iterative Refinement of Tests and Patches for Reliable Software Issue Resolution](https://mlsanity.com/paper/2511.16004/) (2025)
- [MAVUL: Multi-Agent Vulnerability Detection via Contextual Reasoning and Interactive Refinement](https://mlsanity.com/paper/2510.00317/) (2025)
- [Reducing False Positives in Static Bug Detection with LLMs: An Empirical Study in Industry](https://mlsanity.com/paper/2601.18844/) (2026)
- [Sifting the Noise: A Comparative Study of LLM Agents in Vulnerability False Positive Filtering](https://mlsanity.com/paper/2601.22952/) (2026)
- [Cross-Context Review: Improving LLM Output Quality by Separating Production and Review Sessions](https://mlsanity.com/paper/2603.12123/) (2026)
- [More Rounds, More Noise: Why Multi-Turn Review Fails to Improve Cross-Context Verification](https://mlsanity.com/paper/2603.16244/) (2026)
- [A Content-Based Framework for Cybersecurity Refusal Decisions in Large Language Models](https://mlsanity.com/paper/2602.15689/) (2026)
- [AEGIS: From Clues to Verdicts -- Graph-Guided Deep Vulnerability Reasoning via Dialectics and Meta-Auditing](https://mlsanity.com/paper/2603.20637/) (2026)

## Abstract (arXiv)

LLM-assisted defect discovery has a precision crisis: plausible-but-wrong reports overwhelm maintainers and degrade credibility for real findings. We present Refute-or-Promote, an inference-time reliability pattern combining Stratified Context Hunting (SCH) for candidate generation, adversarial kill mandates, context asymmetry, and a Cross-Model Critic (CMC). Adversarial agents attempt to disprove candidates at each promotion gate; cold-start reviewers are intended to reduce anchoring cascades; cross-family review can catch correlated blind spots that same-family review misses. Over a 31-day campaign across 7 targets (security libraries, the ISO C++ standard, major compilers), the pipeline killed roughly 79% of 171 candidates before advancing to disclosure (retrospective aggregate); on a consolidated-protocol subset (lcms2, wolfSSL; n=30), the prospective kill rate was 83%. Outcomes: 4 CVEs (3 public, 1 embargoed); LWG 4549 accepted to the C++ working paper; 5 merged C++ editorial PRs; 3 compiler conformance bugs; 8 merged security-related fixes without CVE; an RFC 9000 errata filed under committee review; and 1+ FIPS 140-3 normative compliance issues under coordinated disclosure -- all evaluated by external acceptance, not benchmarks. The most instructive failure: ten dedicated reviewers unanimously endorsed a non-existent Bleichenbacher padding oracle in OpenSSL's CMS module; it was killed only by a single empirical test, motivating the mandatory empirical gate. No vulnerability was discovered autonomously; the contribution is external structure that filters LLM agents' persistent false positives. As a preliminary transfer test beyond defect discovery, a simplified cross-family critique variant also solved five previously unsolved SymPy instances on SWE-bench Verified and one SWE-rebench hard task.