Introduction
AI chatbots such as OpenAI’s ChatGPT excel at answering general questions. Users enter a prompt, and the chatbot responds instantly based on its existing knowledge.
However, these tools have limitations. They :
- Do not access private systems
- Do not connect to internal applications
- Do not retrieve security or scan data
- Do not process confidential organizational information
- Are not capable of executing any organization specific Automation tasks
For example, if you request, “Show me all Falcon AI issues in my organization,” or “Create a Bit bucket repository for this project”, the AI cannot provide a specific answer or does not know how to execute a request and will either guess or return a generic response.
This is where MCP comes in.
What is MCP?
Before discussing Falcon AI’s MCP, it is important to define what MCP is.
Imagine you are a Falcon AI customer. Over time, you scan multiple applications and collect various issues across different systems and environments.All this information is stored within the Falcon database .
If you use OpenAI or another standalone AI system to ask questions about the data stored in Falcon instance, it cannot assist because it lacks access to your specific context, data, and environment. It can only provide general information, not operational details.
This gap is exactly what MCP fills.
MCP stands for Model Context Protocol.It is a framework that allows AI models to securely access and interpret external, application-specific context such as Falcon AI’s data. This enables them to deliver accurate, relevant, and actionable responses.
How MCP works
Model Context Protocol (MCP) connects your AI assistant to your Falcon AI systems. Unlike tools such as GitHub Copilot, which rely solely on built-in knowledge, MCP enables access to your organization’s private applications, scans, and issue data.
With MCP, assistants gain real context. When a developer asks a question in the IDE, such as fetching issues for an organization or scanning a project, the request is routed through the MCP server for a tailored response.
When a developer asks a question inside the IDE, the request is routed through the MCP server, which then:
- connects to Falcon system securely
- runs the appropriate tools or queries or automation tasks.
- retrieves live results
- sends the response back to the assistant
Responses appear directly in the development environment, enabling teams to query data, identify vulnerabilities, and act without leaving their workflow.
Installing & Enabling the MCP Server
First, users should enable the MCP server within Falcon AI. After logging in to the Falcon AI Suite, navigate to Settings and select the MCP option. The MCP server will then start automatically in the background.
Next, configure your IDE (e.g., VS Code) to connect to the MCP server. Add the provided MCP configuration details to enable the AI assistant or copilot to communicate with Falcon AI.
After completing this one-time setup:
- The IDE will recognize the MCP server.
- The assistant can access Falcon AI tools.
- Queries, scans and automation tasks will work directly within the development environment.
After these steps, all features will work seamlessly within your coding workflow.
Developer Use Cases (Inside the IDE)
Once the MCP server is set up and connected, developers can use Falcon AI right in their coding environment, so there’s no need to switch to the Falcon Web for accessing information. With an AI assistant like GitHub Copilot, users can ask questions or initiate actions in plain language and receive results from the system. From the IDE, developers can:
- Query issues across the organization
Request all issues related to a specific application or organization and get organized results right away. - Fetch project-specific problems while coding
View issues linked to the current project, along with file names and line numbers, so you can fix them right away. - Scan code on demand
Start a Falcon AI Scan from the IDE to analyze your project and automatically upload the results. - Generate specifications (like RAML APIs)
Ask the assistant to create API specifications for you, without having to set everything up yourself. - Slice and dice information conversationally
Ask follow-up questions and narrow down results, all without switching between different dashboards. - Query/Execute any other automation task defined in the Falcon System
MCP Vs Falcon AI
MCP is an integration layer that connects AI assistants to the Falcon system, enabling users to query issues, retrieve scan results, use Falcon AI tools and any other automation tasks directly from their IDE. Any Falcon customer who installs the MCP server can access these features, regardless of whether they have purchased Falcon AI.
Falcon AI, on the other hand, adds generative intelligence, providing advanced features such as
- Automated code generation
- API specification creation
- Test case production
- Automated documentation generation
Real Workflow Example (Day-in-the-Life)
A developer begins work in their IDE and interacts with an AI assistant, such as GitHub Copilot, directly within the editor, eliminating the need to open the Falcon AI dashboard separately.
While coding, developers can request all issues related to their organization or a specific application and receive immediate results. The assistant displays issues, often mapped to specific files and line numbers, enabling prompt resolution.
For deeper analysis, developers can initiate a scan directly from the IDE. The MCP server runs Falcon AI Scan, checks the project, and automatically uploads results, eliminating manual steps.
Customers using Falcon AI can go further by generating API specifications and other artifacts with prompts, all without leaving the coding environment.
This workflow allows users to write code, query issues, scan, fix, and continue in one place, eliminating the need to switch between tools or dashboards.
Conclusion
The key point is that developers can use features of Falcon, Falcon AI and any other Automation tasks internal to the organization without leaving their IDE. MCP integrates all these features directly into the coding environment, eliminating the need to switch between dashboards or tools.
Connecting AI assistants to Falcon systems or any other automation tasks through MCP allows teams to query issues, run scans, and address results directly within their workflows. This integration streamlines processes and improves efficiency.
