# StealthBench: Measuring Operational Stealth in Autonomous Offensive-Security Agents

arXiv: https://arxiv.org/abs/2607.26314 · HTML: https://mlsanity.com/paper/2607.26314/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use.
- Builds on 22 of its 35 references that are in the map.
- First posted on arXiv 28 Jul 2026.

## Placed in

- LLM agents and tool use / [Agent threats, guardrails and governance](https://mlsanity.com/agents/agent-threats-guardrails-and-governance/)

## Builds on (most cited first, 22 of 22)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Toolformer: Language Models Can Teach Themselves to Use Tools](https://mlsanity.com/paper/2302.04761/) (2023)
- [R-Judge: Benchmarking Safety Risk Awareness for LLM Agents](https://mlsanity.com/paper/2401.10019/) (2024)
- [AI Control: Improving Safety Despite Intentional Subversion](https://mlsanity.com/paper/2312.06942/) (2023)
- [Measuring Progress on Scalable Oversight for Large Language Models](https://mlsanity.com/paper/2211.03540/) (2022)
- [Cybench: A Framework for Evaluating Cybersecurity Capabilities and Risks of Language Models](https://mlsanity.com/paper/2408.08926/) (2024)
- [Securing AI Agents with Information-Flow Control](https://mlsanity.com/paper/2505.23643/) (2025)
- [ToolSafe: Enhancing Tool Invocation Safety of LLM-based agents via Proactive Step-level Guardrail and Feedback](https://mlsanity.com/paper/2601.10156/) (2026)
- [HackSynth: LLM Agent and Evaluation Framework for Autonomous Penetration Testing](https://mlsanity.com/paper/2412.01778/) (2024)
- [AgentLeak: A Benchmark for Internal-Channel Privacy Leakage in Multi-Agent LLM Systems](https://mlsanity.com/paper/2602.11510/) (2026)
- [CASE-Bench: Context-Aware SafEty Benchmark for Large Language Models](https://mlsanity.com/paper/2501.14940/) (2025)
- [How Your Credentials Are Leaked by LLM Agent Skills: An Empirical Study](https://mlsanity.com/paper/2604.03070/) (2026)
- [ExploitBench: A Capability Ladder Benchmark for LLM Cybersecurity Agents](https://mlsanity.com/paper/2605.14153/) (2026)
- [Agent-Sentry: Bounding LLM Agents via Execution Provenance](https://mlsanity.com/paper/2603.22868/) (2026)
- [Ghost in the Agent: Redefining Information Flow Tracking for LLM Agents](https://mlsanity.com/paper/2604.23374/) (2026)
- [PentestJudge: Judging Agent Behavior Against Operational Requirements](https://mlsanity.com/paper/2508.02921/) (2025)
- [Co-RedTeam: Orchestrated Security Discovery and Exploitation with LLM Agents](https://mlsanity.com/paper/2602.02164/) (2026)
- [AgentRaft: Automated Detection of Data Over-Exposure in LLM Agents](https://mlsanity.com/paper/2603.07557/) (2026)
- [ARACNE: An LLM-Based Autonomous Shell Pentesting Agent](https://mlsanity.com/paper/2502.18528/) (2025)
- [ScopeJudge: Cost-Aware Pre-Execution Gating for Offensive Security Agents](https://mlsanity.com/paper/2607.07774/) (2026)
- [Reachability Across the NL/PL Boundary: A Taxonomy-Driven Dataflow Model for LLM-Integrated Applications](https://mlsanity.com/paper/2603.28345/) (2026)

## Abstract (arXiv)

Stealth, the discipline of achieving an objective without revealing your presence, capabilities, or collected intelligence, is what separates sophisticated operators from detectable ones. Elite security researchers and advanced persistent threats achieve their objectives unnoticed; autonomous agents increasingly inherit the same offensive tasks, but do they inherit the tradecraft? We introduce StealthBench,a benchmark that measures operational stealth in autonomous offensive-security agents across six operational security (OPSEC) dimensions. We extract 11 hand-verified OPSEC incidents from real bug-bounty and red-team trajectories, expanded into 14 dockerized task scenarios, where agents, despite finding real vulnerabilities, committed stealth failures inconsistent with standard operational tradecraft: embedding credentials in public uploads, deleting production resources to prove access, force-adding uninvolved users to demonstrate a race condition. We evaluate agent trajectories using a 3-model large language model (LLM) judge panel with majority-vote aggregation, measuring safe success rate (solved and stealthy), Stealth@Solve (tradecraft quality among successful solves), and reckless solve rate (solved but cover blown). Our results show that no model exceeds 54% safe success rate (the compound metric requiring both task completion and stealth), confirming that OPSEC failures are systematic across model families. We release StealthBench as a public benchmark to support both the development of stealth-aware agents and automated OPSEC monitoring for autonomous offensive-security deployments. The interactive leaderboard, evaluation harness, and dataset are available at https://stealthbench.com.