Read-only by design
The PowerShell collector gathers a deliberately limited evidence set. It does not repair Windows, retrieve secrets or send reports anywhere.
Support engineering case study
Windows Support Diagnostic Toolkit
A local, read-only diagnostic workflow that collects bounded Windows evidence, validates a structured report and turns it into clear support findings in a Python dashboard.
View repository
- Environment
- Windows / local only
- Repository
- Public
- Focus
- Safe diagnostic evidence
Overview / 01
What the project demonstrates
The project separates data collection, report validation, deterministic evaluation and presentation into distinct stages. That makes the workflow easier to test and prevents the dashboard from silently changing a report or running system commands.
Synthetic fixtures cover healthy, warning, problem, partial and malformed states. Real reports remain local because they can contain machine-specific information, while the public repository uses fictional data for demonstrations and automated checks.
Try it / 02
A simple first run
Use the included fictional sample first. You only need Windows, Python 3.10 or later and a modern browser.
You will need
- Windows PowerShell 5.1
- Python 3.10+
- Git or a downloaded repository ZIP
- 01
Open the project folder
Download or clone the repository, then open PowerShell inside the folder.
- 02
Create the local Python environment
Create and activate a virtual environment so the project dependencies stay isolated.
python -m venv .venv .\.venv\Scripts\Activate.ps1 - 03
Install the requirements
Install the small set of packages used by the dashboard and its tests.
python -m pip install -r requirements-dev.txt - 04
Start the dashboard
The default command opens the fictional sample report. Visit the local address shown below and stop it with Ctrl+C.
python -m dashboard http://127.0.0.1:5000
Thinking / 03
Decisions behind the work
Deterministic support findings
Documented rules produce Healthy, Warning, Problem or Unavailable states, with evidence and plain-English next actions rather than an opaque diagnosis.
Failure-aware presentation
Missing files, malformed JSON, unsupported versions and partial collection produce useful error states instead of raw tracebacks or misleading success screens.
Evidence / 04
Skills and capabilities
- Read-only collection by design
- Healthy, malformed, partial and problem fixtures
- Privacy and failure hardening
- PowerShell
- Python
- JSON
- Schema validation
- Safety testing