PatchGate: a review-readiness gate for GitHub pull requests

[daichunghy]

PatchGate is an open-source CLI and GitHub Action candidate for one narrow question:

Has a pull request supplied the trusted policy, commit-bound evidence, ownership and human boundaries required before a maintainer spends review time?

It is designed to complement GitHub Rulesets, branch protection, CODEOWNERS and CI. It is not an AI-authorship detector, a code-correctness oracle, or a replacement for maintainer judgment.

The current public project includes:

• trusted-base policy loading and discovery-only guidance
• deterministic evaluation and machine-readable ContributionReceipt output
• check and workflow evidence bound to the tested commit SHA
• path ownership, sensitive-path human gates and reviewability signals
• local preflight and a non-blocking shadow workflow

The project is still early. It is 0.1.0-dev, has no Marketplace release yet, and live integration plus external pilots are still open. I am looking for maintainers and platform/security engineers to review the idea and challenge the boundaries.

Useful feedback would be:

1. Which PR failure would make this useful in your repository?
2. Which result would you distrust or find too noisy?
3. What evidence would you require before trying a non-blocking shadow setup?

Repository: https://github.com/daichunghy/patchgate

A critical response is more useful than a vote. Please share your repository shape or a concrete governance edge case if you would like the project to address it.