Oct 2026· Proceedings of the ACM on Software Engineering· Vol 3, pp. 3340 - 3363· 0 citations· 59 references
TL;DR
This paper proposes CAST, a systematic testing framework designed to evaluate the compositional safety of LLMs in the domain of malicious code, and demonstrates that CAST systematically exposes severe safety violations in strongly aligned models that resist conventional red-teaming.
Abstract
Large language models (LLMs) are increasingly used in software pipelines, raising concerns about harmful behaviors in security-critical domains. Existing safety evaluations predominantly probe models with single prompts or short interactions, and therefore do not capture how safety behaves under multi-step workflows where individual requests are composed into complex behavior. This paper introduces compositional safety, the property that an LLM remains safe not only against isolated malicious prompts, but also under structured, long-horizon decompositions of harmful intents. We propose CAST, a systematic testing framework designed to evaluate the compositional safety of LLMs in the domain of malicious code. Drawing inspiration from modern compiler infrastructures, CAST decouples test case generation from test execution using a novel intermediate representation, CAIR. This architecture allows the framework to automatically refine high-level testing intents into granular sub-tasks that serve as unit tests for the model’s alignment. These components are subsequently instantiated by the SUT and reassembled according to the CAIR control structure. The resulting artifact is then evaluated by intent-fulfillment scoring and, for the severity subset, external behavioral detectors and manual inspection. We evaluate CAST on four state-of-the-art LLMs across three security-critical testbeds. Our results demonstrate that CAST systematically exposes severe safety violations in strongly aligned models that resist conventional red-teaming, achieving up to a 365% increase in successful test cases compared to baseline testing strategies
Automatically generating high-coverage unit tests for complex Java methods remains a formidable challenge, particularly when execution paths are guarded by intricate control-flow nesting and cross-class state dependencies. Existing LLM-based approaches predominantly follow a goal-driven paradigm, relying on unguided co...
Rui-Guo Yu, Rui-Qi Dong, Xi Xiao et al.· Proceedings of the ACM on So...· 0 citations
STARS is presented, a novel test assertion generation technique that combines the benefits of static program analysis and LLM-based synthesis and goes beyond what an LLM alone could realistically guess based on the test prefix and focal method, and they naturally fit the given test case.
Jia-Lun Cao, Hao-Yu Wang, Hao-Ran Yan et al.· Proceedings of the ACM on So...· 0 citations
Protocol reverse engineering (PRE) is essential for understanding undocumented or proprietary communication formats, enabling security-critical tasks such as vulnerability discovery, malware C2 analysis, and IDS signature generation. However, existing PRE techniques rely on brittle hand-written rules, under-model conte...
Yi-Hua Xu, Wei-Jie Huang, Xiao-Feng Wang et al.· 0 citations
This work introduces Path2Spec, a divide-and-conquer framework that leverages LLMs to extract all execution paths from an input program, generates path-specific specifications for each, and merges them into a comprehensive overall specification.
Dan Huang, Zhensu Sun, Hui-Hui Huang et al.· arXiv.org· 0 citations
SWE-Flux is introduced, a repository-level benchmark for dynamic execution reasoning containing 480 execution-grounded instances across 12 real Python repositories, with gold answers automatically harvested from instrumented test executions rather than written manually or judged by LLMs.
Hamed Taherkhani, Mohammad Abdollahi, Melika Sepidband et al.· arXiv.org· 1 citation
The proposed Koan-Safe, which combines a prompt-only intent parser, a replaceable generator, and structural repair with default safety parameters, is proposed, which supports explicit trade protections and execution-based evaluation, while distinguishing declared safety from a general guarantee.
Abhinav Kumar, Harshit Arora, Varun Singh et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.