Anthropic Code Review Dispatches Agent Teams to Catch the Bugs That Skim Reads Miss - DevOps.com

"Code Review dispatches a team of AI agents on every pull request to find the bugs that quick reads miss. It's the system Anthropic has been running on nearly every internal PR for months. Now it's available to customers. When a PR opens in an enabled repository, Code Review spins up multiple specialized agents that work in parallel. Some probes for data-handling errors, off-by-one conditions, and API misuse. Others perform cross-file consistency checks and reason about intent."

"Results appear directly on the PR as a single summary comment with inline notes on specific lines. Each finding includes step-by-step reasoning, an analysis of the potential impact, and a suggested fix. Issues are labeled by severity using color codes. The agents do not approve pull requests. Humans decide what to do about the findings."

"After deploying Code Review internally, substantive review comments on PRs jumped from 16% to 54%. Engineers disagreed with fewer than 1% of surfaced findings. Find rates scale with PR size. Changesets over 1,000 lines showed findings 84% of the time. Small PRs under 50 lines had findings 31% of the time."

Anthropic developed Code Review, an AI-powered system addressing the gap between increased code output and inadequate review capacity. With code output per engineer rising 200% annually, only 16% of pull requests received substantive review comments. Code Review deploys multiple specialized AI agents in parallel on each pull request to identify bugs, data-handling errors, API misuse, and consistency issues. The system performs verification steps to filter false positives and ranks findings by severity with suggested fixes. After internal deployment, substantive review comments increased to 54%, with engineers disagreeing with fewer than 1% of findings. The tool deliberately focuses on logical errors rather than style preferences, based on developer feedback.

Read at DevOps.com