Newest papers

Many Repairs, One Direction: Steering Code Models Toward Safer Cryptographic API Use

Tianhe Lu, Sakuna Harinda Jayasundara, Spero Eric, Ewan D. Tempero, Robert Biddle, Giovanni Russello · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.25133 · cited by 0 works (OpenAlex) · open-access PDF

Arabic

في حين أن العديد من الوسائل الموجودة تحسين أمان الكمبيوتر عن طريق تعديل المعايير النموذجية أو الدخول، كيف يمكن أن تتحمل التعديلات التعديلية المتعلقة بتعديلات API التعديلات التعديلية المختلفة تدخلًا مشتركًا لإنتاج الكمبيوتر الأكثر أمانًا لا يزال غير مكتملًا. لبحث هذا، ونحن نتمكن من إنشاء أربعة آلاف زوج من البرمجيات المتوافقة على أربع لغات البرمجيات وأربعين فئة من التطبيقات التعديلية. كل زوج يتغير قرارًا خاصًا بتعديلات API مع الحفاظ على مهام البرمجيات المتوافقة. ونحن نقدم CrypSteer، الذي يتم استخراج إطار مشترك عن طريق قياس

Machine translation (Meta M2M100); the English original follows.

Abstract

Cryptographic API misuse in large language model-generated code can compromise software security. While many existing approaches improve code security by modifying model parameters or inputs, how representation changes associated with different cryptographic API repairs can support a shared intervention for safer code generation remains underexplored. To investigate this, we construct 48,000 matched program pairs across four programming languages and ten cryptographic API misuse categories. Each pair changes a specified API decision while preserving the programming task. Subtracting the misuse program's internal representation from its secure counterpart's gives a repair vector. Across five models, comparisons of mean repair vectors reveal stronger organization by misuse category than by language. These category-specific patterns also support a shared intervention. We introduce CrypSteer, which extracts a shared direction by normalizing repair differences and selecting a principal component aligned with repairs across misuse categories. With model weights fixed, adding one shared direction per model to its internal activations during generation improves aggregate target security rates by 5.28--15.00 percentage points over unsteered generation across languages and misuse categories. These findings connect the shared structure of cryptographic repair representations to a reusable intervention for safer code generation.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: SAFER, Computer science, Cryptographic primitive, Cryptography, Code (set theory), Component (thermodynamics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange