Original | Odaily Planet Daily (@OdailyChina)
Author | Golem (@web 3_golem)
As domestic tech giants rush to launch "one-click OpenClaw installation," controversy has followed.
On March 12, OpenClaw founder Peter Steinberger publicly questioned Tencent's creation of Skillhub on X, accusing it of causing official speed reductions that prevent rapid data scraping, and stating, "they copy but do not support the project in any way."
In response to the controversy, Tencent quickly replied, expressing understanding of Peter Steinberger's concerns. They stated that SkillHub is a localized Skills platform built by Tencent based on the OpenClaw ecosystem. As a local mirror site, it always credits ClawHub as the data source. In its first week, it handled 180GB of traffic for users (870,000 downloads) and only pulled 1GB of non-concurrent requests from the official source. Tencent also expressed willingness to become a sponsor.
Logically, Tencent's response should have clarified the most contentious issue of "whether they are excessively consuming the source site's resources." However, Peter was not satisfied, stating that this was not the main point. He suggested that SkillHub could be made the official fifth mirror to synchronize download statistics, but Tencent should have proactively communicated with him beforehand.
Although the matter ended here, viewing it simply as "the OpenClaw founder's emotional outburst" or "a misunderstanding of a big tech company's normal localization efforts" would be a superficial understanding.
The Problem Isn't the Mirror, But the Big Tech's "High-Handedness"
If you only look at the technical actions, this is not unusual.
In China's developer ecosystem, mirroring open-source projects is a common practice. International open-source infrastructures like npm, PyPI, and Docker Hub have numerous local mirrors in China. Precisely because of this, Tencent denied that its creation of Skillhub was copying, but rather a localized Skills platform. It explained that it was not free-riding or draining the official site but was distributing, accelerating, and adapting to help OpenClaw land in China.
In a sense, Tencent's approach does address the most practical needs of Chinese "lobster farmers." OpenClaw is incredibly popular in China, but not everyone is willing or able to stably access the original community, let alone the fact that the installation, discovery, and retrieval experience for many Skills is still quite primitive.
Skillhub
But is a mirror site inherently blameless? The answer is not necessarily.
Because what the open-source license allows, what community ethics accept, and what commercial reality ultimately brings are often three different sets of accounts.
At the license level, as long as the license is followed and the source is credited, many mirroring and redistribution actions are valid. At the community ethics level, Tencent's SkillHub credits the OpenClaw official source and has actively reduced the source site's bandwidth costs, seemingly taking responsibility.
But Tencent forgot that OpenClaw is not a small open-source project needing big tech's deliberate resource injection. It is the hottest project on GitHub with the most stars. In this context, Tencent's action without prior notice becomes "high-handed." It's no longer just a simple mirroring issue but quickly involves three more sensitive questions: Who represents the official ecosystem? Who is taking the user entry point? And who is defining the download, distribution, and statistical metrics?
This is what truly discomforted Peter, who stated that Tencent's behavior directly affects download statistics. Peter isn't opposed to Tencent localizing OpenClaw for China but believes they should have communicated first, rather than building the platform, onboarding users, and then explaining under public pressure that they were actually there to help.
Furthermore, from a commercial reality perspective, once a platform shell like SkillHub gains scale, the official status and statistical authority originally held by the OpenClaw community can easily be marginalized. Today it's a localized Skills platform; tomorrow it could be the "default Skills distribution market"; later, it might be "who decides which Skills are seen, installed, and commercialized."
This is the real danger signal behind this controversy and a scene all too familiar in the Chinese internet over the past decade: the enclosure movement.
Big Tech Isn't "Farming Lobsters"; They're Using Lobsters to Enclose AI Land
In the past period, "farming lobsters" has become the hottest meme in the Chinese AI circle, and OpenClaw has been rapidly pushed into an almost emotional industry symbol. Everyone says lobsters represent the new imagination of the Agent era and the future of personal AI assistants, which sounds very passionate.
But big tech doesn't look at lobsters with idealism; they see entry points, traffic, distribution power, and the shell of the next-generation operating system.
In the early hours of March 11, Pony Ma promoted Tencent's full suite of "lobster" products on his WeChat Moments. Tencent's "Lobster Family Bucket" customizes a "little lobster" for general users, developers, and enterprise users, supporting one-click installation with no threshold. SkillHub was launched simultaneously at this time, built with 13,000 localized Skills for one-click invocation, usable directly in scenarios like Xiaohongshu operations and Baidu search.
Of course, Tencent isn't the only one "smelling the wind and moving." If you stretch the timeline, you'll find that domestic tech giants have almost collectively jumped in to help users solve the "lobster farming"难题, their movements so整齐 they seem to have pressed the same switch, though Tencent is currently the most comprehensive.
On the surface, everyone means well, but in reality, this hides the most familiar path dependency in the commercial playbook of Chinese internet companies. Facing a new ecosystem already validated by the market and heated up by public opinion, the first move isn't about profit and business models, but grabbing the entry point first, building the platform first, onboarding users first.
What Tencent wants is not just to make it easier for Chinese users to "farm lobsters," but to ensure that when Chinese users truly start "using Agents to get things done," their first instinct is to do it within Tencent's product shell.
This is the most intriguing aspect of actions like SkillHub. It表面上 (superficially) is a mirror site, but in essence, it could be the starting point of a larger closed loop. Today users see local Skill search and download; tomorrow it might be default access to some cloud, some account system, some enterprise workbench. Later, developers might slowly realize that although they are still developing within the OpenClaw ecosystem, the real decisions about exposure, recommendation, review, and commercialization paths are made by the platform.
This script has been played out too many times in the Chinese internet. From ride-hailing to food delivery, from short-video platforms to cloud markets, almost every "ecological prosperity"背后 (behind it) has been accompanied by the same structural ending—the platform first uses free, open strategies to attract users, then builds walls, using traffic, advertising, and other means to turn the ecology back into its own附属层 (subsidiary layer).
Big tech companies all know that old entry points like search, social, content, and e-commerce are fiercely competitive to the limit, while Agent might be the most promising new entry point for the next round. In that case, rather than waiting for OpenClaw to grow wildly on its own, it's better to take it over while it's still in its explosive early stages, encapsulate it first, and get users accustomed to "bossing lobsters around" within their own systems.
Therefore, everyone is too familiar with what will happen next after big tech companies争先恐后 (scramble) to help users solve the OpenClaw installation problem. And Peter, who doesn't understand the Chinese internet, naturally can't comprehend why Tencent didn't communicate with him beforehand or synchronize data with him.
OpenClaw originally represented another AI future: local operation, personal control, community extension, open connection. Its most imaginative aspect is making Agents truly the user's own execution layer. But once this ecosystem gets repackaged by big tech with "localized mirrors," "domestic adaptation," "unified distribution," and "security review," the flavor changes. In the product logic of big tech, the entry point belongs to me, distribution belongs to me, so ultimately, payment and commercialization should also belong to me.
To put it more bluntly, big tech companies aren't "embracing lobsters"; they are "using lobsters to enclose territory in the AI era."
And this is the most unsettling aspect behind this small controversy. Walls are never built all at once; they always grow slowly under the guise of "more convenient" and "more stable." By the time developers, users, and traffic are all contained within the same shell, so-called openness and autonomy might ultimately just become a component within the big tech ecosystem.
OpenClaw currently faces the most paradoxical fate in China: The lobster hasn't fully grown yet, but big tech has already started setting up the nets.
