ReviewSentry v0.3.3-beta — Available Now

ReviewSentry v0.3.3-beta is now live on the GitHub Actions Marketplace. This release adds the most-requested features since the initial beta: a configurable CI gate on verdict, draft PR skip control, colour-coded findings, an extensible criteria configuration system, and a full BDD test suite.

Less time scanning, more time fixing

The promise of ReviewSentry has always been simple: every pull request gets a structured code review before a human looks at it. The AI catches the predictable problems — exposed credentials, merge conflicts, unsafe variable expansion, missing documentation — so your reviewer can focus on the things that actually require human judgement: architecture, business logic, intent.

v0.3.3-beta makes that loop tighter. Reviews now include colour-coded severity indicators (🔴 Critical, 🟠 High, 🟡 Moderate/Low per finding; ✅ / ⚠️ per criterion section), making it faster to triage a long review and identify what actually needs attention. The verdict is now always the absolute last line of the review, so readers engage with the findings before seeing the outcome — which is, after all, the point of reading a review.

What's new

Configurable verdict gate

The new fail_on input lets you configure ReviewSentry as a required status check. Set fail_on: request_changes and the workflow exits non-zero when the AI verdict is REQUEST CHANGES — which means GitHub can block the merge until the issues are resolved. The default remains never, so existing workflows are unaffected.

The verdict is also now exposed as a verdict action output, so you can branch on it in downstream workflow steps.

Draft PR skip control

ReviewSentry reviews draft PRs by default — that is the right behaviour when you want early feedback. But if your team uses drafts for work-in-progress branches that change frequently, you can now set review_drafts: false to hold off until the PR is marked ready for review.

Extensible criteria configuration

Add a .github/reviewsentry.yml file to your repository to customise which criteria apply — without touching your workflow file. Disable criteria that are not relevant to your project (cross-platform checks for a single-target codebase, bash quality for a Python project), and add custom criteria specific to your stack or conventions. Core criteria can be disabled with explicit acknowledgement.

BDD test suite

ReviewSentry now has a full BDD test suite using pytest-bdd with Gherkin feature files. The feature files serve as both the specification and the acceptance tests — new functionality is specced in Gherkin before the code is written. The suite runs in CI on every PR.

Getting started

If you are already using ReviewSentry, update your pinned SHA to the v0.3.3-beta release SHA — available on the Releases page.

If you are new to ReviewSentry, the product page covers setup, configuration, and the recommended PR workflow in full.

ReviewSentry is MIT licensed, free to use, and works at zero additional cost with your existing GitHub account via GitHub Models.