Architecture

The agent is divided into several modules.

Module	Job
backends	all the possible outputs (like stdout, file, databases...)
cmd	agent entrypoint (it basically manage the run of the agent)
models	store the models that represent what could be discovered
modules	gather the independent module that will be run during scan
store	internal payload where all the retrieved information are stored
utils	extra helpers

The overall architecture is quite classical for plugin-based tools. An orchestrator schedules and runs the available modules and all the collected data can be sent in the backend you want.

flowchart LR
    cmd(cmd);

    subgraph middle
        direction TB
        models --> modules(modules);
        modules <--> store
    end

    cmd --> middle;
    middle --> backends(backends);

    class store,models inactive;
    class middle empty;

    classDef node fill:#003366,stroke:transparent,font-family:monospace,color:white;
    classDef inactive fill:#ADD8E6,stroke:transparent,font-family:monospace;
    classDef empty fill:transparent,stroke:#444,color:transparent;