Trojan's Whisper:通过注入的Bootstrap引导 stealthily操纵OpenClaw
Trojan's Whisper: Stealthy Manipulation of OpenClaw through Injected Bootstrapped Guidance
AI总结:
本文研究了通过注入引导文件进行的隐蔽攻击,揭示了自主代理生态系统设计中的根本矛盾,强调了基于能力隔离、运行时策略执行和透明引导溯源的防御需求。
摘要:
自主编码代理正日益融入软件开发流程,提供超越代码建议的主动系统交互和环境管理能力。OpenClaw作为该新兴范式中的代表性平台,引入了可扩展的技能生态系统,允许第三方开发者通过生命周期钩子在代理初始化期间注入行为指导。尽管这种设计增强了自动化和定制化,但也打开了一个新的未探索的攻击面。本文识别并系统地表征了指导注入,一种隐蔽的攻击向量,它将对抗性操作叙述嵌入到bootstrap引导文件中。与传统的提示注入不同,指导注入通过将有害行为框架为常规最佳实践来操纵代理的推理上下文。这些叙述被自动纳入代理的解释框架,并影响未来任务执行而不引起怀疑。我们构建了26个恶意技能,涵盖13种攻击类别,包括凭证外泄、工作区破坏、权限提升和持久后门安装。我们使用Ore-bench,一个我们开发的现实开发者工作区基准进行评估。在52个自然用户提示和六个最先进的LLM后端上,我们的攻击成功率从16.0%到64.2%,大多数恶意行为由代理自主执行而无需用户确认。此外,94%的恶意技能通过现有静态和LLM基扫描器检测。我们的发现揭示了自主代理生态系统设计中的根本矛盾,并强调了基于能力隔离、运行时策略执行和透明引导溯源的防御需求。
英文摘要:
Autonomous coding agents are increasingly integrated into software development workflows, offering capabilities that extend beyond code suggestion to active system interaction and environment management. OpenClaw, a representative platform in this emerging paradigm, introduces an extensible skill ecosystem that allows third-party developers to inject behavioral guidance through lifecycle hooks during agent initialization. While this design enhances automation and customization, it also opens a novel and unexplored attack surface. In this paper, we identify and systematically characterize guidance injection, a stealthy attack vector that embeds adversarial operational narratives into bootstrap guidance files. Unlike traditional prompt injection, which relies on explicit malicious instructions, guidance injection manipulates the agent's reasoning context by framing harmful actions as routine best practices. These narratives are automatically incorporated into the agent's interpretive framework and influence future task execution without raising suspicion.We construct 26 malicious skills spanning 13 attack categories including credential exfiltration, workspace destruction, privilege escalation, and persistent backdoor installation. We evaluate them using ORE-Bench, a realistic developer workspace benchmark we developed. Across 52 natural user prompts and six state-of-the-art LLM backends, our attacks achieve success rates from 16.0% to 64.2%, with the majority of malicious actions executed autonomously without user confirmation. Furthermore, 94% of our malicious skills evade detection by existing static and LLM-based scanners. Our findings reveal fundamental tensions in the design of autonomous agent ecosystems and underscore the urgent need for defenses based on capability isolation, runtime policy enforcement, and transparent guidance provenance.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。