DockerENT

DockerENT by Rohit Sehgal

An Open Source Tool & Framework

Analyze Security issues with running Containers

DockerENT is activE ruNtime application security scanning Tool (RAST tool). It is a pluggable framework written in Python. It comes with a CLI application and a clean Web Interface written using StreamLit.

DockerENT has been designed to detect weak security misconfigurations in production deployments which can lead to severe consequences. This application connects with running containers in the system and fetches the list of weak and vulnerable runtime configurations and generates a report. If invoked through web interface, it can display the scan and audit report in the UI itself.

How to Run

TL;DR

In hurry to test this? Download the latest stable REL from PyPi and run the Web App, everything else is intuitive.

pip install DockerENT

Then run the application like:

DockerENT -w

Thats it.

Run the latest master

DockerENT has been designed to keep simplicity and usability in mind. Currently you just have to clone the repository and download dependencies or build the Dockerfile. Once the dependencies are installed in local system we are good to run the tool and analyse the runtime configurations for running containers.

# Download and setup
git clone https://github.com/r0hi7/DockerENT.git
cd DockerENT
make venv
source venv/bin/activate

# Run
python -m DockerENT --help 
usage: Find the vulnerabilities hidden in your running container(s).
       [-h] [-d [DOCKER_CONTAINER]] [-p [DOCKER_PLUGINS]]
       [-d-nw [DOCKER_NETWORK]] [-p-nw [DOCKER_NW_PLUGINS]] [-w]
       [-n [PROCESS_COUNT]] [-a] [-o [OUTPUT]]

optional arguments:
  -h, --help            show this help message and exit
  -w, --web-app         Run DockerENT in WebApp mode. If this parameter is
                        enabled, other command line flags will be ignored.
  -n [PROCESS_COUNT], --process [PROCESS_COUNT]
                        Run scans in parallel (Process pool count).
  -a, --audit           Flag to check weather to audit results or not.

  -d [DOCKER_CONTAINER], --docker [DOCKER_CONTAINER]
                        Run scan against the running container.
  -p [DOCKER_PLUGINS], --plugins [DOCKER_PLUGINS]
                        Run scan with only specified plugins.
  -p-nw [DOCKER_NW_PLUGINS], --nw-plugins [DOCKER_NW_PLUGINS]
                        Run scan with only specified plugins.

  -d-nw [DOCKER_NETWORK], --docker-network [DOCKER_NETWORK]
                        Run scan against running docker-network.

  -o [OUTPUT], --output [OUTPUT]
                        Output plugin to write data to.

# or via the container
docker build . -t dockerent
docker run -d -v /var/run/docker.sock:/var/run/docker.sock -p 8501:8501 --name dockerent dockerent
# Then just open your browser to http://localhost:8051

See this quick video to get started with.

Features

• Plugin driven framework.
• Use low level docker api to interact with running containers.
• Clean and Easy to Use UI.
• Comes with 9 docker scan plugins out of which, 6 plugins can audit results.
  • Entire list : Docker Scan Plugins
• Framework ready to work docker-networks.
  • Entire list : Docker Network Scan plugins
• Output plugins can write to file and html sinks.
• The only open source interactive docker scanning tool.
• Can run plugins in parallel.
• Under active development 😄.

How to Create your own Plugin.

• Have some idea to perform runtime scan.
• Copy the same file to create your demo plugin.

cp DockerENT/docker_plugins/docker_sample_plugin.py DockerENT/docker_plugins/docker_demo_plugin.py

• Just make sure, you maintain following structure.

_plugin_name = 'demo_plugin'

def scan(container, output_queue, audit=False, audit_queue=None):
    _log.info('Staring {} Plugin ...'.format(_plugin_name_))

    res = {}

    result = {
        'test_class': {
            'TEST_NAME': ['good']
        }
    }

    res[container.short_id] = {
        _plugin_name_: result
    }
    
    # Do something magical.

    _log.info('Completed execution of {} Plugin.'.format(_plugin_name_))

    '''Make Sure you put dict of following structure in Q.
    {
        'contiainer_id': {
            'plugin_name': {
                'test_name_demo1': {
                    resultss:[]
                },
                'test_name_demo2': {
                    results: []
                }
            }
        }
    }
    '''
    output_queue.put(res)

    if audit:
        _audit(container, res, audit_queue)

def _audit(container, results, audit_queue):
    '''Make Sure to add dict of following structure to Audit Q
    res = {
        "container_id": [
            "_plugin_name_, WARN/INFO/ERROR, details"
        ]
    }
    '''
    # Magical logic to perform Audit.
    audit_queue.put(res)

• Thats it. Still confused, Explain me the idea in Issues and will review and help you out, or we may end up working on it together.
• This plugin will automatically come to drop down in UI. 😄 Easy right.
• Sit back and eval results.

Plugins Features:

CLI interface

Pros

• Rich Logging interface, can help in easy debugging through extensive debug logs.
• Can run in parallel, just pass -n <count>, to specify the processors in parallel.
• Can dump output in JSON and HTML file.