Rovo Dev Code Review in Bitbucket and GitHub | Bitbucket Blitz | Atlassian

Quick Overview

Rovo Dev Code Review automates code review processes across Bitbucket and GitHub by enforcing custom coding standards via a review-agent.md file and verifying acceptance criteria from Jira issues, leading to a 45% reduction in PR cycle time for Atlassian engineers.

Key Points: Rovo Dev Code Review standardizes PR processes by using a repository-level review-agent.md file to enforce custom coding standards. Rovo Dev verifies acceptance criteria (AC) listed in Jira issues by analyzing commit messages and the pull request description. The tool automatically flags violations against defined standards and ACs, providing specific comments directly on the pull request. The demonstration showed Rovo Dev identifying violations like missing logging and hardcoded values in a C++ calculator application. The integration works for both Bitbucket and GitHub repositories, with a dedicated configuration screen for enabling the Code Reviewer agent. Atlassian engineers internally observed a 45% reduction in PR cycle time since adopting Rovo Dev Code Review. The process reduces manual developer effort by automating compliance checks, allowing engineers to focus on higher-level architectural work.

Context: This video, presented as a 'Bitbucket Blitz' session by Atlassian Senior Technical Evangelist Warren Marusiak, demonstrates how to utilize Rovo Dev Code Review, an AI-powered agent designed to accelerate the software development lifecycle. The demonstration focuses on setting up and using Rovo Dev to enforce coding standards and validate acceptance criteria on pull requests in both Bitbucket and GitHub environments, using a sample C++ calculator application as an example.

Detailed Analysis

Warren Marusiak introduces Rovo Dev Code Review as an AI agent to streamline development by handling planning, coding, testing, and automating repetitive work in Bitbucket and GitHub. The setup involves navigating to the Rovo Dev agent configuration page (01:25) where the user selects their Atlassian site and agrees to the terms. Code Reviewer agent activation is then shown for Bitbucket repositories, followed by a demonstration of enabling the agent for a GitHub organization (02:03). Custom coding standards are defined in a repository-specific file named (03:35), which lists rules categorized by area, such as Regulated Data (PII/PHI/PCI), Tenant Isolation, Feature Flags, Logging, Domain-specific business rules, Legacy system migration, Design system/accessibility, and Cloud/Infrastructure policies. The presenter then demonstrates adding a log function to a C++ calculator application, which initially violates several documented standards, including missing logging and hardcoded sensitive data exposure (04:35-10:24). When a pull request is created with these changes, Rovo Dev automatically comments on the PR, pointing out the specific violations based on the rules in the review-agent.md file (10:51-17:44). The tool also checks acceptance criteria defined within the Jira issue description (07:17). The demonstration concludes by showing that after fixing the code locally and pushing the changes, Rovo Dev re-reviews the code, verifying that the passing tests pass and the failing/flaky tests are correctly flagged, ultimately providing an overall successful check mark (10:00-10:05, 11:36). The final key takeaways highlight that standardizing code reviews via review-agent.md, verifying acceptance criteria, and using Rovo Dev reduces PR cycle time by 45% for Atlassian engineers by automating granular checks and freeing up developer time for higher-value work.