Agentic Workflows for Cross-Repo Automation

4 posts · Mar 2025 – Aug 2026

What this trail covers

This feature thread tracks the evolution of using agentic workflows to automate cross-repository tasks. Initially, the focus was on enabling AI to assist in code development and review. This post details the implementation of IssueOps, a methodology that leverages GitHub Issues, GitHub Actions, and pull requests as an interface for automating workflows. This allows for event-driven automation of tasks like CI/CD pipelines, approvals, and deployments directly from GitHub Issues, treating issues as a state machine for workflow management.

All GitHub trails

Aug 2026Turn one giant AI-generated pull request to a reviewable stackThis post introduces and details the implementation of GitHub Stacked Pull Requests as a mechanism to manage AI-generated code changes. It explains how to decompose large features into smaller, reviewable units using the `gh stack` CLI, including initializing stacks, adding layers, and managing dependencies. The post also outlines the agentic workflow for creating and managing these stacks, and describes the reviewer's journey through stacked pull requests, emphasizing a top-down read and bottom-up review approach. It also covers the implications of rebase operations on signed commits.
Jul 2026Better tools made Copilot code review worse. Here's how we actually improved it.This post details the engineering effort to improve Copilot code review by optimizing the instructions for shared code exploration tools (grep, glob, view). It describes how the initial migration to these tools led to a regression in performance and effectiveness, characterized by a 'browsing loop' rather than focused investigation. The core contribution is the redesign of tool instructions to align with a reviewer's workflow: starting from the diff, narrowing the search with grep and glob, and using view only for specific evidence. This involved batching discovery and reads, and refining error recovery strategies. The post emphasizes the use of internal benchmarks to debug agent behavior and validate the effectiveness of these instruction changes, resulting in approximately 20% lower average review costs.
Jul 2026Automating cross-repo documentation with GitHub Agentic WorkflowsThis post details the implementation of GitHub Agentic Workflows for automating cross-repo documentation generation. Key contributions include: a markdown-based workflow definition with YAML frontmatter and English prompts; a secure output handler that materializes agent intent (JSON blob) into GitHub actions via a scoped GitHub App; a deterministic target branch resolver mapping product repo milestones to docs repo release branches; and the use of a draft-only pull request strategy with the feature's SME as the reviewer. It also addresses challenges like prompt budget limitations for large diffs by pre-extracting metadata and the need for a mirrored checkout pattern for cross-repo operations.
Mar 2025IssueOps: Automate CI/CD (and more!) with GitHub Issues and ActionsThis post introduces and details the concept of 'IssueOps,' a methodology for automating workflows using GitHub Issues, GitHub Actions, and pull requests. It frames these workflows as finite-state machines, defining states, events, transitions, guards, and actions. The post provides a practical example of building a team membership request and approval process using this methodology, including code snippets for GitHub Actions triggers based on issue comments. It emphasizes the event-driven, customizable, and auditable nature of IssueOps for streamlining operations within GitHub.
Advertisement