Orient Securities released a research report stating that due to the high flexibility of OpenClaw in permission management, a loss of control could potentially lead to large-scale data leaks or erroneous system commands. The emergence of new scenarios signifies new market opportunities, and the cybersecurity industry is expected to welcome new growth prospects. The advent of new technologies like AI Agents not only creates new demand scenarios but also drives the optimization and iteration of security products or solutions, from which the security industry stands to benefit further. The firm believes that comprehensive security providers and those with a focus on identity security are likely to benefit from the development of AI Agents. Key viewpoints from Orient Securities are as follows:
OpenClaw's popularity raises security concerns. On March 10, the National Computer Network Emergency Response Technical Team/Coordination Center of China (CNCERT) issued a "Risk Alert on the Security Application of OpenClaw," pointing out that OpenClaw has been granted high system permissions. It noted that improper installation and usage have already led to risks such as plugin poisoning and security vulnerabilities, recommending that relevant users implement appropriate security measures. The firm believes that due to OpenClaw's high flexibility in permission management, a loss of control could result in large-scale data breaches or incorrect system operations.
OpenClaw expands the cybersecurity perimeter, bringing Agent security into focus. The surge in popularity of OpenClaw is driving a shift in cybersecurity from traditional perimeter protection towards advanced behavior governance targeting AI Agents. As a framework with system-level permissions capable of autonomously executing Shell commands and file operations, OpenClaw extends the attack surface from mere "prompt injection" to "execution chain hijacking." This complexity necessitates that security providers develop products or solutions adaptable to these new scenarios. Currently, specialized detection targeting the OpenClaw plugin (Skill) supply chain and dynamic sandbox technologies to prevent "indirect prompt injection" have emerged within the industry. The firm views the appearance of these emerging scenarios as indicative of new market space, suggesting the cybersecurity industry may see fresh growth opportunities.
AI Agents are also expected to reshape security operation paradigms. The open-source nature and high extensibility of OpenClaw are propelling cybersecurity automation from the "scripting era" to the "reasoning era." Traditional Security Operations Centers (SOCs) rely on pre-set playbooks to handle alerts, whereas OpenClaw or other AI Agents integrated with large language models can autonomously understand the attack intent behind anomalous traffic and coordinate blocking by invoking tools like firewalls and endpoint defenses across platforms. This AI Agent-based security operation model has the potential to significantly alleviate the pressure from a shortage of specialized talent. Concurrently, security vendors can utilize OpenClaw to build automated red team/blue team platforms, simulating Agent attack paths to uncover defense vulnerabilities. The firm concludes that the emergence of new technologies like AI Agents not only spawns new demand scenarios but also promotes the optimization and iteration of security products or solutions, potentially bringing further benefits to the security industry.
Related subjects include: DBAPP Security (688023.SH, Overweight), SDIC Security & Intelligence (300188.SZ, Overweight), Sangfor Technologies (300454.SZ, Buy), NSFOCUS (300369.SZ, Buy), Topsec Technologies (002212.SZ, Buy), Security Awareness (688201.SH, Not Rated), Venusense (002439.SZ, Not Rated), Geer Software (603232.SH, Buy).
Risk warnings include intensified industry competition and commercialization progress falling short of expectations.