Experimental: manifest validation diagnosis on PRs — feedback welcome

[Demitrius Nelon (denelon)]

Warning

This is experimental automation. Its comments are advisory, may be wrong, and never change the state of a pull request.

What it is

We’re piloting an author-assist workflow that provides specific guidance for selected manifest validation failures:

• Manifest-Validation-Error
• Manifest-Installer-Validation-Error
• Manifest-AppsAndFeaturesVersion-Error

The workflow reads the public validation log and submitted manifests, then may post one comment identifying a concrete problem such as a missing schema header, incorrect filename or package path, named schema violation, inconsistent installer metadata, or overlapping Apps and Features version.

Tracking issue: #412466

What it will and won’t do

• ✅ Posts at most one comment per pull-request head SHA.
• ✅ Comments only when the validation log identifies a concrete condition.
• ✅ Remains silent when the available error is ambiguous.
• ✅ Skips wingetbot-authored pull requests.
• ❌ Never edits manifests, labels, or pull-request state.
• ❌ Never approves, merges, closes, waives, or reruns validation.
• ❌ Never invokes wingetbot.
• ❌ Never downloads or executes installers.
• ❌ Never handles security findings or exposes installer URLs or hashes.

Feedback requested

Please share examples of:

• An incorrect or misleading diagnosis.
• A comment that missed an actionable validation error.
• Advice that was unclear, noisy, or redundant.
• A situation where the workflow should have remained silent.
• Wording that could be clearer or more contributor-friendly.

When possible, include links to the pull request and validation run. Do not post installer binaries, hashes, or suspected-malware links.

Created with GitHub Copilot assistance.