D-CIPHER introduces a dynamic collaborative framework where multiple specialized LLM agents work together to solve complex cybersecurity challenges.

• Employs heterogeneous execution with agents specializing in different security skills
• Implements dynamic planning to adapt strategies based on new information
• Demonstrates superior performance over single-agent approaches in CTF challenges
• Creates an extensible architecture for autonomous security analysis

This research advances offensive security capabilities by enabling AI systems to conduct sophisticated security assessments through collaborative problem-solving, potentially transforming how organizations approach vulnerability discovery and remediation.

D-CIPHER: Dynamic Collaborative Intelligent Agents with Planning and Heterogeneous Execution for Enhanced Reasoning in Offensive Security