The paper experimentally evaluates 12 multi-agent LLM collaboration topologies for software design, finding that structural adversarial prompting and cross-model review are the most effective approach…

The paper empirically evaluates the security quality of LLM-generated code across various prompting methods, finding that while prompting alters the structure of weaknesses, it is insufficient to reli…

The paper proposes a novel '3+1' heterogeneous multi-agent architecture using cloud LLMs and a local verifier to achieve high-accuracy, cost-effective code vulnerability detection, significantly outpe…

The paper introduces CodeGolf Bench, a novel multi-language benchmark using code golf to measure LLMs' ability to generate highly concise and efficient code, showing that reasoning models significantl…

The paper introduces 3DCodeBench, a systematic benchmark and platform for evaluating Vision-Language Model (VLM) agents' ability to generate procedural 3D models from text and images using code.

This study analyzes over 20,000 real-world coding sessions to show that AI coding agents frequently fail users through subtle misalignment, requiring constant manual correction even when major system…

The paper introduces Vertical Integration Bias (VIB) to quantify whether LLMs favor their own provider's ecosystem when generating code, finding that this bias is significant in both direct and agenti…

This paper identifies the 'Format-Reliability Gap'—where LLMs know about code vulnerabilities but generate insecure code anyway—and proposes a localized, per-vulnerability steering vector fix that sig…

The paper re-evaluates LLM agents on CTFs, finding that while general-purpose agents like claude-code are strong baselines, specialized, modular architectures significantly improve performance and con…

The paper introduces a data-centric optimization pipeline to improve coding agents' ability to interact with a branching lakehouse, showing significant accuracy gains by treating agent evaluation as a…

The paper introduces MAAD, a multi-agent framework that autonomously transforms software requirements into comprehensive, multi-view architectural blueprints, significantly improving completeness and…

Minor, single-character perturbations to prompts can significantly degrade the security of code generated by LLMs, suggesting that prompt fragility is a major security concern beyond simple prompt inj…

This paper quantifies the polymorphic capacity of a commercial LLM, demonstrating that it can cheaply generate large populations of structurally diverse, yet behaviorally equivalent, offensive code pa…

The paper reports on penetration tests conducted on proprietary, large-scale AI agent systems, finding that security vulnerabilities persist despite stricter development standards.

The paper proposes an automated, standardized framework to empirically compare the security quality of code generated through human-only, LLM-only, and hybrid collaboration methods.

This paper empirically evaluates the security of code generated by seven popular LLMs and finds that all evaluated models generate code containing critical or high-severity vulnerabilities.

The paper investigates how AI coding assistants shift developers' security focus from proactive prevention to reactive review, finding that this structural change is reinforced by current tool interac…

The paper establishes an information-theoretic upper bound on the combined functional capacity and perturbation retention of code LLMs, quantifying the security budget available for code generation.

FPMoE introduces a sparse Mixture-of-Experts (MoE) architecture to improve functional code generation across multiple functional programming languages, achieving state-of-the-art performance with fewe…

The paper introduces TRAILS~, a novel method that improves code correctness validation by grounding LLM reasoning in concrete (input, output) pairs derived from specifications, achieving state-of-the-…