ResearchResearch paperLarge Language Models · Efficiency & Inference1 source · Oct 21, 2026

AutoFraming: Synthesizing Dynamic Frames of C Code with LLMs

To address this limitation, we propose an LLM-based approach for synthesizing ACSL frame specifications alongside functional contracts from C code.

Large language models (LLMs) have shown considerable promise in synthesizing formal specifications from source code.
However, existing approaches primarily focus on functional contracts and loop invariants, while largely neglecting frame conditions, which specify the memory locations that may or may not be modified during execution and constitute an essential part of formal specifications.
Our approach analyzes code property graphs (CPGs) to extract frame-relevant information from the input code and leverages this information to guide the LLM in generating accurate frame conditions.
On a benchmark of 84 ACSL-compliant functions from non-trivial real-world systems, our method successfully synthesizes error-free frame specifications for 92.9% of the functions, representing a 179% relative improvement over state-of-the-art techniques.

[1]AutoFraming: Synthesizing Dynamic Frames of C Code with LLMs
arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 21, 01:22 AM
To address this limitation, we propose an LLM-based approach for synthesizing ACSL frame specifications alongside functional contracts from C code.
Large language models (LLMs) have shown considerable promise in synthesizing formal specifications from source code.

Extractive summary: sentences quoted from the sources.

Before this