[HeyItsGilbert]

Summary

.github/
└── instructions/
    └── articles.instructions.md   # applyTo: content/articles/**/*.md
GLOSSARY.md                        # new: AI slop, vibe coded, off-topic pitch

Path-specific Copilot custom instructions, scoped to community article submissions only (layers on top of the existing repo-wide .github/copilot-instructions.md, which both review surfaces already apply alongside it). Copilot checks three categories on every article PR:

review(diff)
  if unsafe: destructive commands / credential harvesting / prompt injection / phishing links / CoC violations
    comment(quote, category)
  if AI slop: generic filler / no firsthand specificity / untested-looking code /
              re-explained docs / listicle padding / no PR validation evidence
    comment(quote, category)
  if off-topic: thin PowerShell content relative to self-promotion
    comment(quote, category)
  always: one triage summary comment — "Needs maintainer safety review: <category>"
  never: imply the PR is blocked or rejected — maintainer has final say

Built from:

• a grilling session settling scope, categories, and enforcement stance (advisory-only, since Copilot code review can't block merges)
• the precedent in #265, where "vibe coded"/off-topic pitch content was caught manually
• the last 9 accepted article PRs (Add article: Validate Azure Resource Relationships with PSRule and PowerShell Graphs #57–Build Interactive Vega Visualizations in Verso with PowerShell #377), which surfaced the "Validation" section convention (tool versions, commands run, links checked) as a concrete, checkable signal — now folded into the AI-slop checklist

Evidence

• applyTo: "content/articles/**/*.md" glob verified against the live tree: matches 1,210 files, correctly including both article bodies and _index.md listing pages (per explicit scoping call), no false matches outside content/articles/.
• Frontmatter parses cleanly (applyTo is the only key, no YAML errors).
• No permanent tests apply — this is a prose instructions file read by GitHub's Copilot review surfaces, not executable code.

Merge Danger

Door: Two-way — it's a markdown instructions file; edit or delete anytime with no migration cost.

Blast Radius: Low. Only affects Copilot's own review comments on PRs touching content/articles/**/*.md; no build, routing, or runtime behavior changes. Worst case is a false-positive advisory comment, which a maintainer can dismiss — Copilot has no merge/block authority.

[netlify]

✅ Deploy Preview for powershellorg ready!

Name	Link
🔨 Latest commit	a0523f4
🔍 Latest deploy log	https://app.netlify.com/projects/powershellorg/deploys/6ae9323f361226819457c630
😎 Deploy Preview	https://deploy-preview-519--powershellorg.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.
🤖 Make changes	Run an agent on this branch

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[Copilot]

🟡 Changes recommended

The validation-evidence check can falsely flag issue-form PRs whose generated descriptions omit submitter evidence, so it should check the article or linked submission before flagging missing validation.

1 open issue

• Check article and submission for validation evidence before flagging · New

What changed in this PR

Adds path-specific Copilot review guidance for community articles, extending the repository-wide instructions without changing site behavior.

Changes:

• Defines checks for unsafe content, low-quality AI-generated content, and off-topic promotion.
• Requires specific evidence and a triage summary while leaving merge decisions to maintainers.
• Adds glossary definitions for the review terminology.

File	Description
GLOSSARY.md	Defines AI slop, vibe coded, and off-topic pitch.
.github/​instructions/​articles.instructions.md	Adds advisory review checks scoped to article Markdown files.

_{Review effort: Balanced}