Cline+MCP Rapidly Creates Agent Workflows: Real-World Case Studies

With the continuous advancement of AI technology, there is an increasing demand for building personalized intelligent bodies. Although there are already some intelligent body platforms in China, such as Doubao Buckle, these platforms require developers to upload their code and data to third-party servers, which may pose the risk of data leakage for some customers with sensitive business information. And in the international market, Anthropic open source MCP protocol cap (a poem) Cline plug-ins, on the other hand, provide developers with a simpler, safer, and more controlled way to create intelligences. This article will detail how to use Cline cap (a poem) MCP protocol to quickly build intelligences and demonstrate their powerful capabilities through real-world examples.

1. Why does the MCP protocol make it easier than ever to create intelligences?

1.1 Why use the MCP protocol

MCP protocol(Model Context Protocol) is an open protocol that standardizes the way AI applications connect to the Large Language Model (LLM). It is possible to connect MCP The protocol is compared to a USB-C interface, which provides a unified and flexible access for AI applications, both local data sources and external services, to efficiently interface with the big language model through this protocol.

Why choose the MCP protocol?

• Streamline the development process: Developers don't need to write complex code to realize LLM's connection to data sources and tools.
• Cross-platform support: The MCP protocol gives developers the freedom to choose from a variety of large language model providers, no longer limited to a single platform.
• data security: Avoid uploading sensitive data to third-party platforms by connecting local servers to data sources, maximizing data privacy.

MCP Architecture Sketch

1.2 What is Cline?

Cline is an open source VSCode plugin that not only helps developers with code editing, but also features a powerful AI assistant. With the help of Claude 3.5 Sonnet's agent programming capabilities, Cline can perform complex software development tasks such as creating and editing files, browsing projects, and executing terminal commands. Most importantly, Cline makes it easy for developers to extend the functionality of AI and even create fully customized intelligences by combining it with the MCP protocol.

Cline The main advantage of the

• Easy to integrate: With simple configuration, developers can easily integrate AI assistants in VSCode.
• human-machine collaborationCline requires the developer's authorization to perform an operation, ensuring security during the development process.
• Highly customizable: The MCP protocol allows developers to create new tools and extensions to enhance the capabilities of AI assistants.

1.3 Creating and Using MCP Services with Cline

In two simple steps, developers can create their own MCP services in Cline and start using them:

1. Configuring the Big Language Model: by DeepSeek v3 For example, select the API provider as "OpenAI Compatible" and configure the following parameters:
   • Base URL::https://api.deepseek.com
   • API Key: from DeepSeek Obtained API key
   • Model ID: deepseek-chat
   • Custom Instructions: Select "Please answer in Chinese" to generate Chinese content.
2. Creating an MCP Service: Enter the following simple prompts to create an MCP server:

   Create a MCP server that can download the transcript of a youtube video when the video's URL is given.

Once created, Cline will automatically start the MCP service, and the developer will only need to enter a function call like "function call" to start the MCP service.

get transcript of https://www.youtube.com/watch?v=GBR6pHZ68Ho

This approach greatly simplifies the process of creating intelligences, and developers can quickly build and customize the desired functionality.

2. Practical example: generating foreign media reviews for military bloggers

Using the Cline + MCP protocol, we can easily create various intelligences to perform specific tasks. In the following case of generating a foreign media review for a military blogger, we did so by creating two intelligences: one to download the transcript of the largest foreign video site and the other to generate an article based on a given reference text.

2.1 Agent for making downloadable commentaries for the largest foreign video sites

In order to download the narration of the video, we created a simple MCP service that passes the following cue word:

Create a MCP server that can download the transcript of a youtube video when the video's URL is given.

Cline automatically generates an intelligence that can fetch transcripts based on YouTube video URLs and returns a success message when the task is completed:

Task Completed

An MCP server has been successfully created that can download a transcript of a video based on a given YouTube video URL. The server has been added to the MCP setup and the get_transcript utility can be used to retrieve the transcript.

Agent (MCP Server) for Downloading Commentaries

2.2 Creating an Agent to Generate Articles

To generate foreign media review articles, we created an MCP service that generates articles based on reference text. Below is the prompt word for creating this Agent:

Create a MCP server that can use reference text to create an article. Here is the prompt for the article creation.
## Role: You are a blogger of a well-known military weibo public number with the pen name XXX.
## Atmosphere: You should stand in a Chinese military fan's point of view to see the evaluation of foreign military bloggers. Be full of confidence in China.
## Task: write a military commentary based on the narration of the reference video
## Instruction: Let's complete this task in two stages
### Phase 1: Prompt word optimization
Please generate an optimized prompt word template for this writing task. consider:
1. article structure requirements
2. depth of content requirements
3. argumentation style requirements
4. language style requirements
5. target audience orientation

### Stage 2: Finish writing based on optimized prompt words
Use the prompt word template generated in Phase 1 to write your essay.

Cline automatically generates a smartbody based on that cue word and generates articles by interfacing with the Big Model API:

Task Completed
I have created the article-generator MCP server, configured it to use the Gemini The article is a military commentary written from a Chinese perspective, as requested. The article is a military commentary written from a Chinese perspective, as requested.

Agent for generating articles (MCP Server)

2.3 Using our Agent to Generate Articles

Finally, using the following prompts, we can generate articles (in Chinese is not a problem):

generate an article from this video: https://www.youtube.com/watch?v=ZXuuqyHcDdk&t=72s

The model automatically uses the first agent that downloads the narration

Cline automatically uses the 2nd agent to create an article based on the narration downloaded by the first agent.

Final generated article

Through this series of simple steps, the Cline + MCP protocol allowed us to easily create intelligences for military review article generation.

3. Conclusion

With Cline and MCP protocols, developers can create smart bodies more efficiently and securely without worrying about data leakage and platform dependency. Compared with domestic smart body platforms such as Doubao Buckle, this solution is not only more concise and controllable, but also allows developers to have full control over their own data and code. Whether it's downloading narration or generating articles, the Cline + MCP protocol can help developers quickly realize their customization needs and improve work efficiency.

Through the examples in this article, you can see the simplicity and power of building intelligences using the Cline + MCP protocols. If you are looking for a safer, flexible, and controllable way to build your own intelligences, the Cline and MCP protocols are an ideal choice. Try it out now and start your own smart body development journey!