Brokerage Industry Stirred by "Lobster" - Prohibition of Private Installation Yet Frequent Research Reports

When OpenClaw (an open-source AI agent, commonly known as “Lobster”) sparked a “shrimp farming” craze in social circles, the brokerage industry was caught in a highly contrasting “magical realism” game.

On one side, some brokerages have issued strict internal bans on installing “Lobster” on company devices; on the other side, many research institutes view it as a core variable in the new fintech ecosystem. As of now, brokerage firms such as Founder Securities (601901), GF Securities (000776), and others have released special research reports, deeply analyzing the industry transformation and investment opportunities behind it.

This parallel of “ban” and “research” reflects the financial industry’s dual considerations of security compliance and technological innovation.

Some brokerages ban “Lobster” installation internally

“Our company issued a notice internally prohibiting the private installation of ‘Lobster.’”

A brokerage professional working in Beijing told reporters: “The financial industry has nearly strict requirements for data security. The open-source nature of OpenClaw and its high system permissions conflict with the existing risk control systems of brokerages. The notice states that, for risk management reasons, it is strictly forbidden to install or use any version of OpenClaw on company-assigned desktops, laptops, production and testing servers, internal shared storage devices, company-issued work phones, or within the company’s network environment.”

However, this professional also mentioned that employees’ personal phones or computers are not within the company’s network scope, and thus the company does not prohibit installation or use of “Lobster” on personal devices. “Because the company cannot monitor private information, but currently, usage is not encouraged.”

Two other brokerage personnel indicated that their companies have not officially issued restrictions on installing or using “Lobster,” but do not encourage it either. “Regarding technological development, especially AI development, companies generally keep a close watch, but daily emphasis is on employees strictly following the company’s cybersecurity policies to ensure communication environment security.”

Previously, the Ministry of Industry and Information Technology, the National Internet Emergency Center, and others have repeatedly warned about the security risks of OpenClaw applications.

Recently, the Cybersecurity Threat and Vulnerability Information Sharing Platform monitored that some instances of OpenClaw open-source AI agents pose high security risks under default or improper configurations, which could easily lead to network attacks, information leaks, and other security issues.

Because OpenClaw’s deployment involves “blurred trust boundaries” and features such as continuous operation, autonomous decision-making, and system and external resource calls, without effective permission control, auditing mechanisms, and security reinforcement, it may execute unauthorized operations due to command induction, configuration flaws, or malicious takeover, leading to information leaks, system control, and other security risks.

It is recommended that relevant units and users thoroughly check exposure to public networks, permission configurations, and credential management when deploying and using OpenClaw, disable unnecessary public network access, improve security mechanisms such as identity authentication, access control, data encryption, and security auditing, and continuously monitor official security notices and reinforcement suggestions to prevent potential cybersecurity risks.

Some brokerage teams are conducting research

Contrasting sharply with internal bans, many brokerage research teams are accelerating their study of OpenClaw. As of March 11, at least eight brokerages—including Founder Securities, GF Securities, Dongwu Securities (601555), Guojin Securities (600109)—have released dedicated research reports on OpenClaw, covering core areas such as multi-platform deployment, investment research scenario implementation, and reshaping computing power demands. Interestingly, some of these brokerages have internal policies prohibiting the installation of “Lobster.”

Founder Securities released a report at the end of February titled “OpenClaw Empowers Financial Investment Research: 17 Efficient Application Cases Explained,” which states that OpenClaw, as a pioneering open-source personal AI assistant platform, is valuable because it successfully integrates large language model cognition with local system execution capabilities, bridging the gap from “dialogue” to “action.” Recently, OpenClaw has attracted market high attention, reflecting a strong market demand for practical, controllable AI tools that can truly enhance productivity.

The Founder Securities research team conducted detailed tests on scenarios such as daily document management, email management, automated task reminders, AI picture book generation, deep research on agents, and automated browser-based information retrieval and analysis. They stated: “Overall, the efficiency and quality of completion are very high. For professionals engaged in active investment research, OpenClaw can significantly reduce the difficulty of building various tools, data, and quantitative stock-picking strategies, freeing investors from repetitive and rule-based work, allowing them to focus more on complex decision-making and innovative strategy development. For quantitative research professionals, it can be used for factor research, strategy reproduction, portfolio construction, and testing, greatly improving work efficiency.”

Zheshang Securities, in their report “Next-Generation Investment Research Infrastructure: From Deployment to Application of OpenClaw,” explained the role of OpenClaw for different user groups: “It does not replace research personnel but liberates humans from the physical labor of information processing, enabling them to focus on strategy innovation, complex judgment, and value creation. For active investment researchers, it is an efficiency lever; for quantitative researchers, a strategy accelerator; for individual investors, a tool for capability equalization.”

It is worth noting that almost all brokerage research personnel, while embracing OpenClaw, have issued risk warnings.

Zheshang Securities pointed out that the autonomy of OpenClaw may introduce operational risks, requiring strict permission boundary configurations, such as prohibiting automatic trading execution and sensitive information confidentiality. The outputs of large language models are inherently stochastic, and generated content may produce “hallucinations”; key data and information need manual verification. Users should exercise caution, ensuring compliance and legal considerations when collecting data with AI tools.

Dongwu Securities also warned that because OpenClaw has “super permissions” of the operating system, improper configuration or unreliable third-party Skill packages could lead to accidental deletion or leakage of important local files.