GitHub Targets Large Merge Problem with Stacked PRs

2026-04-29 · Source: InfoQ · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning · Depth: Intermediate, short

Summary

GitHub has introduced a native stacked pull request workflow via a new CLI extension, gh-stack, aiming to simplify the review and merging of large pull requests. This approach, where each branch targets the preceding branch in a series rather than the main branch, allows developers to continue work while earlier layers are under review. Research indicates that PRs between 200 and 400 lines have 40% fewer defects and are approved three times faster. The gh-stack extension automates complex mechanics like cascading rebases and force-pushes, and integrates a stack map into GitHub's UI for easier navigation. It also includes an AI agent integration to assist in creating and managing stacks. This move addresses a long-standing need previously met by third-party tools like Graphite, which offers similar features.

Key takeaway

For Machine Learning Engineers managing complex feature development, adopting GitHub's new gh-stack for stacked pull requests can significantly streamline your workflow. This native integration helps break down large changes into smaller, more reviewable units, potentially reducing defects and accelerating merge times. Consider experimenting with the gh-stack CLI and its AI agent integration to improve code quality and team velocity, especially for projects prone to large, conflict-heavy PRs.

Key insights

Stacked PRs improve code review efficiency and quality by breaking large features into smaller, manageable changes.

Principles

• Smaller PRs reduce defects and accelerate approval.
• Dependent branches enable continuous development.
• Automated tooling simplifies complex rebase operations.

Method

Developers create a series of dependent branches, each targeting the previous one, allowing work on later layers while earlier ones are reviewed. gh-stack sync cascades rebases and force-pushes atomically.

In practice

• Use gh-stack to manage dependent pull requests.
• Aim for PRs between 200-400 lines for better outcomes.
• Explore AI agents for stack creation and management.

Topics

• GitHub
• Stacked Pull Requests
• gh-stack
• Code Review Workflow
• AI Agent Integration

Best for: Machine Learning Engineer, Software Engineer, DevOps Engineer, AI Engineer

Related on AIssential

• Rethinking Git for the Age of Coding Agents with GitHub Cofounder Scott Chacon — Git's original design and backward compatibility limit its utility for modern AI agent workflows, necessitating new version control approaches.
• ghapi, a new third-party Python client for the GitHub API — ghapi simplifies GitHub API interaction for Python developers via a comprehensive, OpenAPI-driven library and CLI.
• v2.1.81 — Claude Code updates significantly enhance AI development workflows through new models, expanded context, and improved stability.
• GitHub CEO: Our Devs Barely Write Code — AI can automate code review and vulnerability remediation, freeing developers for more complex tasks.
• Pullfrog AI: Open-Source CodeRabbit Alternative Powered by GitHub Actions — Pullfrog AI offers a model-agnostic, open-source AI agent framework for GitHub, running entirely within GitHub Actions.
• GitHub Slashes Agent Workflow Token Spend up to 62% with Daily Audits and MCP Pruning — GitHub cut LLM agent token spend up to 62% via daily audits, MCP pruning, and GitHub CLI integration.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by InfoQ.