Tue Le, Anh Nguyen Hoang, Gabriel Sasseville, Doseok Jang, Pierre‐Luc Bacon, Nghi D. Q. Bui · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.25916 · cited by 0 works (OpenAlex) · open-access PDF
जबकि कोडिंग मॉडल और एजेंट हार्डवेयर अधिक सक्षम हो जाते हैं, भरोसेमंद सुरक्षा सत्यापन एक केंद्रीय बोतलन बन जाता है। सुरक्षित कोड जनरेटिंग में सुधार करने के लिए एक सिग्नल की आवश्यकता होती है जिसकी रिपोर्ट की प्रगति अनुकूलन के दौरान वास्तविक प्रोग्राम व्यवहार के साथ संरेखित रहती है; अन्यथा, एक ऑप्टिमाइज़र सिग्नल के फैसले को बेहतर कर सकता है, बिना आधारित कमजोरी को हटाने के। मौजूदा मॉडल-आधारित दृष्टिकोणों को परीक्षण के तहत प्रोग्राम के निष्पादन ट्रैक के माध्यम से प्रोग्राम व्यवहार को निष्कर्षित करने के लिए मॉडलों की आवश्यकता होती है. परीक्षण समय में मरम्मत के लिए सही परिणामों को निर्दिष्ट करने के लिए, हम
Machine translation (Meta M2M100); the English original follows.
As coding models and agent harnesses become more capable, reliable security verification is becoming a central bottleneck. Improving secure code generation requires a signal whose reported progress remains aligned with actual program behavior under optimization; otherwise, an optimizer can improve the signal's verdict without removing the underlying vulnerability. Existing model-based approaches require models to either infer program behavior from source code or specify correct outcomes for adversarial tests, placing substantial reasoning demands on the security signal. We introduce \falcon{}, a counterexample-grounded verification framework that separates attack exploration from behavioral verification and connects them through execution traces of the program under test. In test-time repair, \falcon{} yields substantially stronger held-out functional-security gains than alternative security signals across secure code generation benchmarks, while preserving or improving functionality. As an RL reward, \falcon{} also outperforms static-analysis and learned rewards on held-out functional-security, even when all policies successfully optimize their respective training rewards. Combining \falcon{} with either proxy reward further improves held-out performance, suggesting that execution-grounded feedback can mitigate reward overoptimization while preserving useful guidance from existing security signals.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Computer science, Adversarial system, Coding (social sciences), Source code, Code (set theory), Computer security
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.25916, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange