The tech giants are converging on another intensely competitive new frontier: the desktop agent. On February 6th, the official account for "Tencent Cloud CodeBuddy" announced the commencement of internal testing for its desktop agent tool, "WorkBuddy," marking the entry of another major contender into this closely watched "desktop examination." The Tencent Cloud CodeBuddy team stated that WorkBuddy was launched internally at Tencent some time ago, with the aim of helping non-developer employees within the company improve their daily efficiency. Observing a growing demand among office workers, researchers, and various industries for "applications to solve daily tasks," the team leveraged CodeBuddy's mature AI Coding infrastructure to launch a product characterized by a lightweight, convenient, and minimalist design philosophy. "We will continue to deeply refine the product based on user feedback, aiming for an official launch at an appropriate time in the first half of the year," the team added. Unlike the currently popular OpenClaw, which can invoke system-level permissions and perform more flexible operations, WorkBuddy is positioned more as a tool that "enables various autonomous operations by reading authorized folders on the computer." This implies that WorkBuddy's use cases are more focused on limited operations within clearly defined folder constraints. This perhaps reflects Tencent's "subtractive" product thinking in this domain. It assumes that most professionals care not only about "getting the job done" but also worry about AI causing problems. Consequently, the chosen product path is not aggressive, instead prioritizing stable, reusable, and deliverable execution results. The launch of WorkBuddy not only signifies Tencent's formal entry into the desktop agent tool market competition but can also be seen as a significant signal that major tech firms are attempting to redefine the boundaries of workplace efficiency through the new variable of "intelligent agents." Tencent's timing coincides with the global desktop agent market reaching a tipping point. Shortly before this, products like OpenClaw gained rapid popularity in technical communities, with users amazed by AI's ability to deeply access computer systems and files, even autonomously executing tasks across different applications. This experience, dubbed the "24/7 on-call Jarvis," led the industry to declare that the "ChatGPT moment for desktop agents" has arrived. Compared to other popular tools currently on the market, WorkBuddy does not follow a purely geek-oriented route. Instead, it is positioned for workplace professionals without technical backgrounds. This means it no longer requires users to possess programming knowledge; they simply need to describe their needs in natural language, and it will execute tasks like a "knowledgeable" colleague. This paradigm shift from "technical tool" to "workplace assistant" represents a bold attempt by Tencent to popularize desktop agents. According to reports, WorkBuddy can autonomously plan and execute multi-modal complex tasks directly on a local computer. Specifically, WorkBuddy possesses two main characteristics: First is its deep local operation capability: it can "understand natural language" and "actually operate local files." After receiving authorization, it can read folders, batch process files, generate documents or PPTs, perform deep data analysis, conduct industry research, and handle multiple agent tasks in parallel. This means it breaks the traditional AI limitation of being "able to see but not act," truly addressing the last mile of office automation. Second is the autonomous planning of multi-step tasks. Unlike simple command responses, WorkBuddy emphasizes "thinking with a brain." Furthermore, WorkBuddy comes built-in with various domestic and international models, mainstream MCP Servers, and Skills packages, among other advanced features. This gives it strong extensibility while maintaining ease of use, allowing it to meet the deep needs of vertical scenarios like finance, research, and content creation. It is understood that WorkBuddy was developed by the Tencent Cloud CodeBuddy team. It was previously launched internally at Tencent for testing, with over 2,000 employees from different roles participating in deep usage. It has seen widespread application in scenarios such as data processing and analysis, building local knowledge bases, content and copywriting creation, poster generation, and office automation. For a long time, although large language models appeared omniscient within chat interfaces, they often seemed "restricted" when users returned to real office scenarios, facing complex local files and numerous task switches. Now, as AI competition enters its later stages, the industry's focus is shifting from "competing on model parameters" to "competing on implementation capabilities." Simple chat bots can no longer meet the complex, fragmented, and highly local-file-dependent needs of professionals. Whoever occupies the user's computer desktop first holds the ticket to the next generation of "AI operating systems." The popularity of tools like OpenClaw has already demonstrated users' intense demand for agents that "can run on their own computers and deeply operate applications." Tencent's move at this time aims both to secure a niche amidst a field of pioneers and to intelligently upgrade its deep-rooted enterprise service capabilities through the vehicle of the Agent. However, behind the reignited market excitement, WorkBuddy and other desktop agent players also face challenges. Foremost among these is the trade-off between "overreach" and "privacy." The core value of a desktop agent lies in its deep access to the user's computer system, files, and applications. This depth of interaction means users must grant very high system permissions. Just as the industry expects an "on-call Jarvis" from OpenClaw, the more powerful the Agent, the more detailed the personal preferences and local project data it possesses. Ensuring convenience while preventing the misuse or leakage of this sensitive data during network transmission will be a critical compliance "line in the sand" that AI Agent products must address. The Tencent Cloud CodeBuddy team stated that WorkBuddy relies on CodeBuddy's security protection mechanisms and employs a multi-layered defense strategy of "sandbox isolation + Skill standardization + dangerous operation blocking" to address the challenges of desktop Agent deployment. Its core philosophy is: "Authorize trust and verify" – trusting the intent of user instructions, verifying execution safety through user authorization, and having a blacklist for dangerous operations as a safety net. Furthermore, the "parent-child relationship" between the Agent and its underlying foundation model determines the upper and lower limits of its capabilities. Although desktop Agents can optimize execution paths through engineering methods, their fundamental capabilities are still derived from the underlying large model. If the base model experiences "hallucinations" or logical breaks, the Agent could potentially cause serious consequences during autonomous task execution. More importantly, the rise of desktop agents essentially redefines the interaction between humans and computers. Historically, users acted as "carriers" between applications, moving between browsers, office suites, and communication tools. With the proliferation of desktop agents like WorkBuddy, the future interaction paradigm may shift to being "task-centric." When an Agent can cross applications and complete complex tasks on behalf of the user, such as batch unsubscribing from emails or organizing price discrepancies, traditional application software might gradually become mere "skill packages" invoked by Agents. This suggests that future desktop ecosystem competition may no longer be about stacking software features, but about whose interfaces are more conducive to Agent invocation and who can provide an operating environment better suited for AI. Globally, the competition for desktop agents has just begun. Beyond names like Manus, OpenClaw, and Claude Coworker, numerous other similar products are rushing to launch. When AI truly takes control of our mouse and keyboard, it represents the ultimate enhancement of efficiency, but also a new test of digital security boundaries. As the "desktop agent war" intensifies, a transformation in productivity paradigms is already at our doorstep.