CocoLoop: The Largest OpenClaw AI Agent Skills Store in China 162

CocoLoop is a skill store and community built around OpenClaw (commonly known as "小龙虾" / "Crayfish") in China, launched by the Molili platform under Dangbei ai. It primarily addresses the problems of slow access to overseas skill repositories, lack of security guarantees, and poor localization.

The skill store is at hub.cocoloop.cn, and the community is at www.cocoloop.cn. It currently includes over 5,000 Skills, covering scenarios such as web automation, document processing, development assistance, multimodal generation, and API calls, and is compatible with more than 20 mainstream AI Agent tools.

Main Features of CocoLoop

• Skill Distribution & Installation: The platform supports three methods: one-click web installation, CLI batch deployment, and offline manual import. Each Skill comes with detailed usage guides. For example, the multimodal summary assistant requires you to configure your own API Key, and the community has corresponding troubleshooting tutorials.

• Domestic Mirror Acceleration: Self-built CDN mirror nodes provide much faster download speeds than overseas sources. A 50MB skill package might take 20 minutes to download from an overseas repository, while CocoLoop's domestic mirror handles it in seconds.

• Security Review: Every Skill listed must pass CLS/BSS full-chain security detection, receive an S/A/B/C/D level risk rating, have its permission scope labeled, and have malicious code blocked. Enterprise users can also use a VM-level isolated execution environment.

• Curated Rankings & Topics: There's a Top 50 curated Skills list, selected manually and with secondary security authentication, covering scenarios such as information retrieval, autonomous evolution, and collective memory. There are also some scenario-based topic packs, such as the "Crayfish Farming 101" beginner must-install pack, OpenClaw investment Skills collection, and top Skills pack for entrepreneurs. These come with configuration tutorials and work out of the box.

• Chinese Community: This is currently the most popular OpenClaw discussion area in China, covering content such as deployment tutorials, Skill development, model reviews, error troubleshooting, and practical AI monetization. Users can post questions, share self-made skills, and exchange real-world experience.

CocoLoop Application Scenarios

• Office Automation: Document processing, spreadsheet parsing, PDF conversion, report organization

• Content Creation: Xiaohongshu viral title generation, copywriting imitation, multimodal summarization, video script assistance

• Data Collection: Compliant crawler for fetching public data

• Cross-border E-commerce: Product information organization, price comparison automation, batch listing

• Development Assistance: API calls, code processing, automation scripts, test case generation

• Investment Research: Stock monitoring, new stock subscription reminders, NFT market tracking, research report analysis

• Personal Productivity: Schedule organization, information aggregation, morning briefing robots, timed tasks.

Pros and Cons of CocoLoop

Pros:

Fast domestic access without needing a proxy; security review is stricter than overseas sources, reducing malicious code risks; active Chinese community where you can find people to ask when you run into problems; curated rankings and topic packs are beginner-friendly, so you don't have to search for a needle in a haystack; compatible with all versions of OpenClaw and 20+ Agent platforms.

Cons:

The core ecosystem is tied to OpenClaw. If you use other Agent tools such as Claude Code or Cursor, some Skills have limited compatibility; many Skills require you to prepare your own API Key, which still creates a barrier for people with no technical background at all; community content quality varies, with some posts being fluff or ads; the platform itself does not directly provide AI conversation capabilities and needs to be used with clients such as Molili or ChatClaw.

How to Use CocoLoop

Prepare the Environment

First, install an OpenClaw-type client. Molili (Dangbei's Chinese version) is recommended—it's ready to use out of the box and doesn't require you to configure an API. Windows users should make sure the installation path doesn't contain Chinese characters. macOS users encountering "Cannot verify developer" should go to System Settings and click "Open Anyway".

Access the Skill Store

Open hub.cocoloop.cn, browse the Top 50 rankings or scenario topics, and find the Skill you need.

Install a Skill

Each page provides a dedicated installation command, generally in the following format:

npx skills install [skill-name]

Run it in the OpenClaw terminal and the domestic mirror will automatically accelerate.

Configuration & Usage

Some Skills require additional configuration (such as filling in an API Key or setting up a Docker sandbox). Just follow the page documentation. After installation, use OpenClaw's built-in commands to view, uninstall, or update the Skill.

Join the Community

If you encounter errors or want to share your experience, post in the corresponding section of www.cocoloop.cn. Common problems (such as "rate limit exceeded" or model integration failures) generally have ready-made solutions in the community.

CocoLoop vs. Competitors

Why Choose CocoLoop

If you are a domestic OpenClaw user looking for a skill store with fast downloads, security screening, and Chinese community support, CocoLoop is currently the most hassle-free option. You don't need a proxy like with overseas sources, and the barrier isn't as high as with pure enterprise platforms. Beginners can follow the Top 50 rankings to install Skills, while experienced users can hunt for custom skills in the community—basically covering most AI Agent extension needs.

FAQ

Q: Does CocoLoop charge a fee?

Skills themselves are free to download and install. However, some Skills that call third-party APIs (such as OpenAI or Claude) require you to prepare your own API Key, and that portion is billed based on usage on the third-party platform. The Molili client itself has a pay-as-you-go top-up model.

Q: Can I use it without OpenClaw?

Q: What should I do if Skill installation fails?

First check whether your Node.js version is 18 or above, then confirm that the installation path contains no Chinese characters or spaces. If it still reports an error, search the community for the error keywords—most common issues already have answers.

Q: What is the relationship between CocoLoop and Molili?

Molili is the Chinese version of the OpenClaw client released by Dangbei AI, while CocoLoop is its配套 skill store and community. Molili is the phone, and CocoLoop is the app store + forum.

Q: Are the third-party one-click installation packages in the community reliable?

Community packaging tools like "秋芝" are used by quite a few people, but their version updates lag behind the official releases, and their security is not as well-guaranteed as official channels. It's fine to try them out for initial experience, but for long-term, in-depth use, going through the official installation is recommended.