Search papers, labs, and topics across Lattice.
This paper introduces MACGen, a multi-agent framework designed to enhance the generation of secure and functionally correct code by integrating planning, security analysis, and code synthesis. By employing specialized roles for each agent鈥攑lanner, security advisor, coder, and reviewer鈥擬ACGen effectively reduces context bloat and improves the quality of code outputs. The framework demonstrates significant performance improvements, achieving an average increase of 19.61 and 10.57 percentage points in functional and security correctness on the CWEval and BaxBench benchmarks, respectively.
Multi-agent collaboration in code generation can boost functional and security correctness by over 19 percentage points, reshaping how we approach secure coding.
Despite their strong ability to generate code, large language models often fail to produce secure code, as their outputs frequently contain security vulnerabilities. Secure code generation is inherently challenging because it requires solving a multi-objective problem: functional correctness and security. Existing approaches address this challenge by injecting external security knowledge or by using agentic feedback and iterative refinement. However, guideline retrieval often leaves the generator to translate generic advice into task-specific secure implementations, while shared-dialogue multi-agent feedback can blur role boundaries and suffer from context bloat. We present MACGen, a multi-agent framework that integrates planning, security analysis, code synthesis and refinement to jointly optimize security and functionality. A planner constructs a step-by-step plan to satisfy functional requirements. A security advisor identifies likely CWEs and synthesizes task-specific guidelines, a coder then generates code grounded in these artifacts, and a reviewer issues perspective-separated feedback. Rather than sharing full dialogue histories, each agent receives only structured artifacts from upstream stages, enforcing role specialization and reducing uncontrolled context growth. On CWEval and BaxBench, MACGen improves F&S@1 over direct prompting by 19.61 and 10.57 percentage points (pp) on average, respectively.
