Skip to main content

Why QZX

Give your agent
a clear starting point.

QZX brings common diagnostics into documented commands and clear results. The same briefing helps you and the automation working alongside you.

Free and open sourceWindows, Linux and macOS support varies by commandHuman output and JSON

When the task repeats

Reuse the inspection. Keep the context.

This diagram shows how the work changes when a QZX operation covers the task. It is a conceptual workflow comparison.

Before

Rebuild the inspection

  1. Choose the command

    Identify the utilities and arguments available in the environment.

  2. Interpret the output

    Gather the data returned by each check.

  3. Prepare the context

    Organize the observations before deciding what to do next.

With QZX

Start from a documented operation

qzx diagnoseProject . --json
  1. Request a diagnosis

    Use the command that supports the task.

  2. Read the evidence

    Review technologies, dependencies, findings and checks that have not run.

  3. Choose the next step

    Share the same result with a person or an agent.

Captured example from a test project

Technology
Python
Unique packages
1
Findings
1
View JSON excerpt
{
    "success": true,
    "details": {
        "technologies": [
            "Python"
        ],
        "dependencies": {
            "unique_declared_package_count": 1,
            "unique_packages": [
                "httpx"
            ]
        },
        "version_control": {
            "status": "not_repository"
        },
        "file_scan": {
            "scanned_file_count": 3,
            "scan_complete": true
        },
        "summary": {
            "issue_count": 1,
            "verification": {
                "release_readiness": "not_assessed"
            }
        },
        "validation": {
            "tests": {
                "status": "configured_not_run"
            },
            "lint": {
                "status": "configured_not_run"
            }
        }
    }
}

What the interface brings

Less repeated work around each task.

A shared vocabulary and results you can inspect.

One documented operation

On supported systems, reuse the same command name and argument model. The catalog explains each operation's parameters and requirements.

Explore the commands

Readable for you. Structured for your agent.

Human-readable output is the default. Add --json when a program or agent needs structured data from the same result.

Read the result contract

Evidence to guide the next step

Execution status and findings occupy separate fields. You can recognize a completed check whose result still needs attention.

See real examples

Where it fits

Choose the tool for the work.

QZX works alongside APIs and native commands. Use it when it covers the operation and its shared interface adds clarity.

When it helps

  • You need a documented diagnosis that works on your supported systems.
  • You want to give an agent a structured result with useful context.
  • You prefer to review the evidence before running scripts or changing files.

What you still need to check

  • Availability, dependencies and platform differences in the command reference.
  • Which checks actually ran: discovering a test does not establish that it passed.
  • Permissions, backups and previews for operations that change state. QZX uses your permissions; a shared interface does not isolate execution.

For platform-specific capabilities, streaming or a stricter typed integration, choose the appropriate native API or tool.

Start with a concrete task

See the result before adding it to your workflow.

Explore the captures and try a read-only command in your own environment.

Where these results come from

Captured on 2026-10-08 with QZX 0.2.2.0.15, CPython 3.13.13 and Windows 11 10.0.26200.

These figures belong to the recorded examples. Copyable commands run in your terminal and their results depend on the environment.