Repository navigation
[Article] AI Can Write the PowerShell. Who Proves It Should Run? #265
Description
Activity
Hello, just following up on this pitch. I would welcome any feedback on whether the topic would be a good fit for PowerShell.org, or whether you would prefer a narrower practical focus. Happy to shape the outline around your guidance.
Thanks,
TDMy initial reaction is that this feels like a pitch for your site. That is not a hard block, but the content really needs to be about PowerShell the language and it's use. Abstract ideas and general approach are totally fine and feel free to reference your site, but the content should really contribute to the use and understanding of PowerShell. If the article feels like it was just vibe coded/written, then we may not post it. A good rule of thumb, if you're not writing at least 75% then it you may get push back.
@HeyItsGilbert: Thanks, that makes sense.
The intention wasn’t to make this a site or framework pitch, so I’m happy to tighten the scope and make the article much more PowerShell-led. I’ll focus it around practical PowerShell patterns for assessing and safely automating an operational task, with concrete examples around evidence collection, validation, decision logic, controlled execution and verification.
The AI angle will stay secondary and be used mainly to discuss reviewing generated PowerShell rather than becoming the subject of the article itself. Any external reference will be kept minimal and relevant.
I’ll also make sure the article is authored from my own engineering experience and not presented as AI-generated content. Thanks for the steer.
Reacted by Gilbert SanchezSounds good 😄 I look forward to reading it!
Article Title
AI Can Write the PowerShell. Who Proves It Should Run?
Author Name
TD
Submission Type
Pitch — I'd like feedback before writing
Description
Generative AI can produce useful PowerShell remarkably quickly, but working code is only one part of safe operational automation. This article follows a practical infrastructure example from evidence collection through interpretation, implementation, controlled execution and verification.
Category
PowerShell for Admins
Tags (optional)
PowerShell, Automation, AI, Testing, Operations
Summary / Pitch
Generative AI has dramatically lowered the barrier to creating PowerShell. An engineer can describe a problem, generate candidate code and have something executing within minutes.
That is useful, particularly for people developing their PowerShell skills.
It also moves the difficult part of automation somewhere else.
The important questions increasingly become:
I would like to explore those questions through a practical infrastructure example rather than a purely conceptual AI discussion.
The article would begin with a read-only operational question and progressively develop the PowerShell around it:
Question → Evidence → Interpretation → Code → Action → Verification
AI would be used during the development process for activities such as exploration, explanation, candidate implementation and test generation, but it would not be treated as the authority deciding whether an operational change is justified.
The PowerShell examples would demonstrate patterns such as structured evidence collection, explicit health or configuration rules, defensive error handling, dry-run or bounded execution, logging and post-action validation.
A central theme would be the difference between:
“The script completed successfully.”
and:
“The intended operational state was successfully established.”
The goal is to give newer PowerShell users a way to think beyond syntax while also giving experienced administrators a reusable model for reviewing automation that may increasingly be written or accelerated with AI.
The article would remain vendor-neutral and focused on practical PowerShell engineering.
Article Content (Markdown)
Author Website or Social Link (optional)
https://stateworks.co.uk
Submission Agreement