Skip to content

Repo Reviews quickstart＃

Learn how to run system-wide codebase analysis with kluster.ai Repo Reviews. Connect your repository, wait for the deep scan to complete, and review cross-module bugs that slip through PR-level reviews.

Prerequisites＃

Before getting started, ensure you have:

  • A kluster.ai account: Sign up on the kluster.ai platform if you don't have one.
  • A connected repository: You'll need to connect your GitHub or GitLab repository through the dashboard.

How repo reviews work＃

Repo reviews analyze your entire codebase as a system instead of reviewing individual changes. When multiple parts of your code interact, issues emerge that don't show up in PR-level reviews.

  1. Connect your repository: Link your GitHub or GitLab repository to kluster.ai.
  2. Scan starts automatically: Wait for the deep analysis to complete.
  3. Review the findings: Examine issues grouped by severity and type.

This catches cross-module bugs, state inconsistencies, bypassed validation checks, and other system-wide problems that slip through regular code reviews.

Running your first repo review＃

1. Connect your repository＃

Go to the Repo Reviews dashboard and click Connect Repository.

Repo Reviews dashboard with Connect Repository button

Select your Git provider (GitHub or GitLab):

Select Git provider modal

Then choose the repository you want to analyze:

Select repository from dropdown

2. Scan repo＃

Once your repository is connected, the analysis starts automatically and shows "Review in progress...". The scan runs a deep analysis of your codebase. Depending on repository size, this takes several minutes. You'll receive an email notification once it's done.

You can close the page and come back later.

3. Review the results＃

When the scan finishes, you'll see a list of issues found in your codebase:

Repo review results showing list of issues

Each issue displays:

  • Description: Summary of the problem.
  • Severity: High, Medium, or Low (shown with color badges).
  • Type: The category of issue (Security, Logical, Performance, Knowledge, etc.).
  • Priority: P0-P5 ranking for triage.

4. Examine issue details＃

Click any issue to see the full analysis:

Issue detail view with description, explanation, and recommended actions

The detail view includes:

  • Description: What the problem is.
  • Explanation: Why this is a problem and how it impacts your system.
  • Recommended Actions: Steps to fix the issue.

5. Take action＃

For each issue, you have three options:

  • Copy: Copy the issue details to share or save.
  • Fix with AI: Get a prompt to paste into your AI assistant (Claude, Cursor, etc.) to fix it automatically.
  • Dismiss: Mark as false positive or accepted risk.

Next steps＃

  • Pick your workflow: Learn when to use repo reviews vs. other review modes.
  • Review modes: Understand all available review types.
  • FAQ: Common questions about kluster.ai code reviews.