Want to learn how to USE AI technology to make money and/or your life easier? Join our FREE AI community here: https://www.skool.com/ai-with-apex/about
China’s Viral ‘Colleague Skill’ Repo Turned AI Anxiety Into a Workplace Flashpoint
A meme-like GitHub project in China landed at exactly the right moment to expose a deeper fear: workers are no longer just worried about AI automating tasks, but about AI absorbing the judgment, tone, and know-how that made them valuable in the first place.
That is why a small open-source repo became a much bigger story about labor, ownership, and the uneasy line between knowledge transfer and knowledge extraction.
TL;DR
- A viral Chinese GitHub project called colleague.skill sparked debate by claiming a coworker’s work style and knowledge could be distilled into a reusable AI skill.
- The repo spread less because of technical novelty than because it touched a nerve around layoffs, AI adoption, and white-collar insecurity.
- The backlash quickly expanded into questions about consent, privacy, and whether job-specific expertise belongs to workers or employers.
- The story unfolded during a wider China AI-agent boom, with businesses and institutions pushing adoption while security concerns also rose.
- The episode matters beyond China because it shows how the AI labor debate is shifting from task automation to the capture of tacit human knowledge.
The repo that made workers feel like datasets
What happened
A project widely referred to as colleague.skill or Colleague Skill went viral in China after claiming it could turn a coworker’s work logic, communication style, and domain knowledge into a reusable AI “skill.” South China Morning Post reported that the developer, Zhou Tianyi, said the project was created in under four hours, but it quickly became a symbol of something much larger than a quick hack.
Why it matters
The repo resonated because it reframed a familiar workplace fear. Instead of asking whether AI can automate part of a job, it raised the prospect that the most human layer of office work could be extracted from chats, documents, and workflow traces, then packaged into software.
Key details
- The project was discussed as a way to “distill” a colleague into a reusable AI skill using workplace artifacts such as chat histories and documents.
- SCMP reported the repo spread rapidly on Chinese social media and GitHub.
- The developer said it was built in under four hours.
- Reactions split across three camps: joke, practical knowledge-transfer tool, and warning sign about worker disposability.
- The strongest public response centered on symbolism, not proof of a formal, widespread corporate practice.
Source links
https://www.scmp.com/tech/tech-trends/article/3349790/colleague-skill-ai-job-fears-china-set-viral-spread-supposed-ability-harvester?utm_source=openai
Why the reaction was so intense
What happened
The project hit a workforce already under pressure. Reporting on China’s AI-driven workplace anxiety has shown growing concern among white-collar workers that AI tools are not only changing jobs, but narrowing the gap between training a system and training a replacement.
Why it matters
This is what gave the story its force. The repo was provocative, but the labor conditions around it made it feel plausible enough to trigger real panic about replacement, surveillance, and the conversion of personal expertise into company assets.
Key details
- Worker anxiety appears tied to broader job insecurity and pressure to adapt to AI-heavy workflows.
- The emotional core of the debate is whether workers are producing output or becoming training material themselves.
- Commentary around the repo focused on tacit knowledge: judgment, habits, tone, and domain instincts that are hard to formalize but highly valuable.
- The story reflects a shift from automation of tasks toward extraction of professional identity and know-how.
Source links
https://restofworld.org/2026/china-ai-anxiety-openclaw-jobs-redundancy/?utm_source=openai
https://aisecwatch.com/issues/aaea0b27-fd17-48b6-b4c9-43562fc9a714?utm_source=openai
https://www.scmp.com/tech/tech-trends/article/3349790/colleague-skill-ai-job-fears-china-set-viral-spread-supposed-ability-harvester?utm_source=openai
China’s AI-agent boom made the story bigger
What happened
The colleague.skill moment arrived during a wider rush into AI agents in China. Reuters reported in March that OpenClaw adoption was growing quickly even as officials and experts raised concerns about how these autonomous systems handle security and sensitive information.
Why it matters
Without that broader boom, the repo might have stayed a niche curiosity. Inside an environment where AI agents are moving rapidly from demos into workflows, the idea of converting a coworker into an on-demand software layer suddenly looked less like satire and more like an early warning.
Key details
- China has been moving aggressively to adopt AI agents across products and workplace systems.
- Reuters reported growing OpenClaw uptake alongside security concerns.
- China also launched an AI literacy push in April 2026, showing official support for broader AI adoption.
- That combination of encouragement and caution helps explain why workplace agent experiments are expanding while scrutiny is rising at the same time.
Source links
https://www.reutersconnect.com/item/china-embraces-openclaw-ai-agent-despite-security-concerns/dGFnOnJldXRlcnMuY29tLDIwMjY6bmV3c21sX1ZBNTA0NTE4MDMyMDI2UlAx?utm_source=openai
https://en.people.cn/n3/2026/0410/c90000-20445476.html?utm_source=openai
https://www.axios.com/2026/03/23/openclaw-agents-nvidia-anthropic-perplexity?utm_source=openai
Backlash, ownership, and the question companies cannot dodge
What happened
Pushback appeared almost immediately. Reports described a counter-tool called anti-distillation.skill, while broader commentary focused on whether an employee’s hard-won know-how should be treated as personal capital rather than as company data available for AI training.
Why it matters
This is where the debate moves from viral spectacle to a durable workplace issue. If companies can capture work traces and use them to build agents, then disputes over consent, access, and ownership will become central to how AI gets deployed inside organizations.
Key details
- A reported counter-project, anti-distillation.skill, emerged as a symbolic response to AI-based knowledge harvesting.
- Commentary cited by Cybernews said some Chinese observers argued job-specific expertise should be treated as a personal asset.
- The core ownership question is whether valuable workplace knowledge belongs to the worker, the employer, or the systems where that work was recorded.
- The safest framing is not that firms are universally cloning employees, but that workers increasingly fear their expertise can be repackaged without meaningful control.
Source links
https://aihola.com/article/colleague-skill-china-ai-clone?utm_source=openai
https://cybernews.com/ai-news/china-ai-clones-employees/?utm_source=openai
https://www.scmp.com/tech/tech-trends/article/3349790/colleague-skill-ai-job-fears-china-set-viral-spread-supposed-ability-harvester?utm_source=openai
The security problem is real too
What happened
Beyond labor concerns, researchers have been warning about the operational risks of powerful agents with broad permissions. Recent academic work on OpenClaw-style systems found that autonomous agents connected to files, email, and internal services can introduce serious security weaknesses.
Why it matters
That makes the “clone your coworker” idea more than just emotionally loaded. Any system built from internal chats, documents, workflows, and broad organizational access creates both a labor problem and a security problem, especially when deployed faster than governance can catch up.
Key details
- Researchers have warned that autonomous agents with broad access can expose sensitive data and workflows.
- Security concerns help explain why Chinese officials and experts have shown caution around some workplace uses of agent systems.
- The more an agent is designed to mirror a colleague’s behavior across tools and services, the higher the potential exposure if controls are weak.
Source links
https://arxiv.org/abs/2604.04759?utm_source=openai
https://www.reutersconnect.com/item/china-embraces-openclaw-ai-agent-despite-security-concerns/dGFnOnJldXRlcnMuY29tLDIwMjY6bmV3c21sX1ZBNTA0NTE4MDMyMDI2UlAx?utm_source=openai
The lasting significance of colleague.skill is not the repo itself. It is the way one viral project captured a new fear of the AI era: that the most valuable part of white-collar work may no longer be the task you perform, but the trace you leave behind for a system to learn from.
—
Want to learn how to USE AI technology to make money and/or your life easier? Join our FREE AI community here: https://www.skool.com/ai-with-apex/about
