README.md•6.43 kB
# DefectDojo MCP Server
[](https://badge.fury.io/py/defectdojo) <!-- Add this badge if/when published to PyPI -->
This project provides a [Model Context Protocol (MCP)](https://github.com/modelcontextprotocol/specification) server implementation for [DefectDojo](https://github.com/DefectDojo/django-DefectDojo), a popular open-source vulnerability management tool. It allows AI agents and other MCP clients to interact with the DefectDojo API programmatically.
## Features
This MCP server exposes tools for managing key DefectDojo entities:
* **Findings:** Fetch, search, create, update status, and add notes.
* **Products:** List available products.
* **Engagements:** List, retrieve details, create, update, and close engagements.
## Installation & Running
There are a couple of ways to run this server:
### Using `uvx` (Recommended)
`uvx` executes Python applications in temporary virtual environments, installing dependencies automatically.
```bash
uvx defectdojo-mcp
```
### Using `pip`
You can install the package into your Python environment using `pip`.
```bash
# Install directly from the cloned source code directory
pip install .
# Or, if the package is published on PyPI
pip install defectdojo-mcp
```
Once installed via pip, run the server using:
```bash
defectdojo-mcp
```
## Configuration
The server requires the following environment variables to connect to your DefectDojo instance:
* `DEFECTDOJO_API_TOKEN` (**required**): Your DefectDojo API token for authentication.
* `DEFECTDOJO_API_BASE` (**required**): The base URL of your DefectDojo instance (e.g., `https://your-defectdojo-instance.com`).
You can configure these in your MCP client's settings file. Here's an example using the `uvx` command:
```json
{
"mcpServers": {
"defectdojo": {
"command": "uvx",
"args": ["defectdojo-mcp"],
"env": {
"DEFECTDOJO_API_TOKEN": "YOUR_API_TOKEN_HERE",
"DEFECTDOJO_API_BASE": "https://your-defectdojo-instance.com"
}
}
}
}
```
If you installed the package using `pip`, the configuration would look like this:
```json
{
"mcpServers": {
"defectdojo": {
"command": "defectdojo-mcp",
"args": [],
"env": {
"DEFECTDOJO_API_TOKEN": "YOUR_API_TOKEN_HERE",
"DEFECTDOJO_API_BASE": "https://your-defectdojo-instance.com"
}
}
}
}
```
## Available Tools
The following tools are available via the MCP interface:
* `get_findings`: Retrieve findings with filtering (product_name, status, severity) and pagination (limit, offset).
* `search_findings`: Search findings using a text query, with filtering and pagination.
* `update_finding_status`: Change the status of a specific finding (e.g., Active, Verified, False Positive).
* `add_finding_note`: Add a textual note to a finding.
* `create_finding`: Create a new finding associated with a test.
* `list_products`: List products with filtering (name, prod_type) and pagination.
* `list_engagements`: List engagements with filtering (product_id, status, name) and pagination.
* `get_engagement`: Get details for a specific engagement by its ID.
* `create_engagement`: Create a new engagement for a product.
* `update_engagement`: Modify details of an existing engagement.
* `close_engagement`: Mark an engagement as completed.
*(See the original README content below for detailed usage examples of each tool)*
## Usage Examples
*(Note: These examples assume an MCP client environment capable of calling `use_mcp_tool`)*
### Get Findings
```python
# Get active, high-severity findings (limit 10)
result = await use_mcp_tool("defectdojo", "get_findings", {
"status": "Active",
"severity": "High",
"limit": 10
})
```
### Search Findings
```python
# Search for findings containing 'SQL Injection'
result = await use_mcp_tool("defectdojo", "search_findings", {
"query": "SQL Injection"
})
```
### Update Finding Status
```python
# Mark finding 123 as Verified
result = await use_mcp_tool("defectdojo", "update_finding_status", {
"finding_id": 123,
"status": "Verified"
})
```
### Add Note to Finding
```python
result = await use_mcp_tool("defectdojo", "add_finding_note", {
"finding_id": 123,
"note": "Confirmed vulnerability on staging server."
})
```
### Create Finding
```python
result = await use_mcp_tool("defectdojo", "create_finding", {
"title": "Reflected XSS in Search Results",
"test_id": 55, # ID of the associated test
"severity": "Medium",
"description": "User input in search is not properly sanitized, leading to XSS.",
"cwe": 79
})
```
### List Products
```python
# List products containing 'Web App' in their name
result = await use_mcp_tool("defectdojo", "list_products", {
"name": "Web App",
"limit": 10
})
```
### List Engagements
```python
# List 'In Progress' engagements for product ID 42
result = await use_mcp_tool("defectdojo", "list_engagements", {
"product_id": 42,
"status": "In Progress"
})
```
### Get Engagement
```python
result = await use_mcp_tool("defectdojo", "get_engagement", {
"engagement_id": 101
})
```
### Create Engagement
```python
result = await use_mcp_tool("defectdojo", "create_engagement", {
"product_id": 42,
"name": "Q2 Security Scan",
"target_start": "2025-04-01",
"target_end": "2025-04-15",
"status": "Not Started"
})
```
### Update Engagement
```python
result = await use_mcp_tool("defectdojo", "update_engagement", {
"engagement_id": 101,
"status": "In Progress",
"description": "Scan initiated."
})
```
### Close Engagement
```python
result = await use_mcp_tool("defectdojo", "close_engagement", {
"engagement_id": 101
})
```
## Development
### Setup
1. Clone the repository.
2. It's recommended to use a virtual environment:
```bash
python -m venv .venv
source .venv/bin/activate # On Windows use `.venv\Scripts\activate`
```
3. Install dependencies, including development dependencies:
```bash
pip install -e ".[dev]"
```
## License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## Contributing
Contributions are welcome! Please feel free to open an issue for bugs, feature requests, or questions. If you'd like to contribute code, please open an issue first to discuss the proposed changes.