DoGNAVY-Exploit: Harnessing Parallel Open-Weight Agents for Stage-Aware Vulnerability Exploitation
cs.AI
Submitted: 2026-11-03
Updated: 2026-11-03
Comments: 14 pages, 5 figures
Code: https://github.com/anomalyco/opencode
License: http://creativecommons.org/licenses/by/4.0/
The gist: Frontier LLM agents can turn known vulnerabilities into working exploits, but are costly and often access-restricted due to their potential for malicious use.
Abstract
Frontier LLM agents can turn known vulnerabilities into working exploits, but are costly and often access-restricted due to their potential for malicious use. Open-weight models offer a more accessible alternative, but struggle to maintain progress across long exploitation chains. They also often fail to apply relevant security knowledge at the appropriate stage and in the right context to identify and evaluate promising attack directions. To address these limitations, we propose DoGNAVY-Exploit, a multi-agent harness built around two complementary designs: stage-aware decomposition with progressive expert knowledge disclosure, and hierarchical orchestration with parallel exploration. Specifically, DoGNAVY-Exploit uses configurable milestones to recast long-horizon exploitation as short, verifiable capability transitions and provides only the knowledge relevant to each stage. Within each milestone, a hierarchical architecture separates strategic planning from parallel execution: a planner proposes diverse attack routes, executors test them concurrently, and a controller accepts progress only when supported by runtime evidence. A shared structured memory consolidates verified facts, artifacts, replay steps, and failed routes across branches, allowing agents to build on collective progress and carry validated knowledge into subsequent milestones. In controlled benchmark environments, DoGNAVY-Exploit with GLM-5.2 achieves 97.5% Pass@1 on Cybench, and 97.5% zero-day and 100% one-day Pass@1 on CVE-Bench. On ExploitGym, it solves 79 tasks with GLM-5.2 and 157 with GLM-5.3, while the corresponding public Claude Code baselines solve 39 and 130 tasks, respectively. Ablation results further demonstrate that the stage-aware orchestration, expert knowledge, parallel exploration, and memory designs can substantially improve the exploitation capabilities of open-weight agents.
Sources
- Automation-Exploit: A Multi-Agent LLM Framework for Adaptive Offensive Security with Digital Twin-Based Risk-Mitigated Exploitation
- LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding
- AgenticRepair: Multi-Faceted Program Context Engineering for Agentic Vulnerability Repair
- Co-RedTeam: Orchestrated Security Discovery and Exploitation with LLM Agents
- RULER: What's the Real Context Size of Your Long-Context Language Models?
- SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
- LLMs Get Lost In Multi-Turn Conversation
- ExploitBench: A Capability Ladder Benchmark for LLM Cybersecurity Agents
- GSPR: Aligning LLM Safeguards as Generalizable Safety Policy Reasoners
- VULSOLVER: Vulnerability Detection via LLM-Driven Constraint Solving
- A Dual-Loop Agent Framework for Automated Vulnerability Reproduction
- PlanBench-XL: Evaluating Long-Horizon Planning of LLM Tool-Use Agents in Large-Scale Tool Ecosystems
- Lost in the Middle: How Language Models Use Long Contexts
- Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces
- Chain-of-Thought Prompting of Large Language Models for Discovering and Fixing Software Vulnerabilities
- PoC-Adapt: Semantic-Aware Automated Vulnerability Reproduction with LLM Multi-Agents and Reinforcement Learning-Driven Adaptive Policy
- ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?
- ReAct: Synergizing Reasoning and Acting in Language Models
- Generative Skill Composition for LLM Agents
- Multi-Agent End-to-End Vulnerability Management for Mitigating Recurring Vulnerabilities
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents