Simplify recently changed code and open low-risk refactor pull requests
This entry turns GitHub Next's Code Simplifier workflow into a narrow cleanup agent. The agent inspects code changed in the last day, proposes behavior-preserving simplifications, runs validation, and opens small refactor pull requests instead of attempting broad rewrites.
npx skills add agentskillexchange/skills --skill simplify-recently-changed-code-and-open-low-risk-refactor-pull-requests
This entry is based on GitHub Next’s Code Simplifier workflow from the githubnext/agentics repository. Upstream, the workflow is installed through gh-aw and scheduled to look only at recently modified code. That narrowness is exactly why it works as an ASE entry. The agent’s job is not “be a general refactoring framework.” Its job is to scan code changed in the last 24 hours, find low-risk simplifications such as flattening nested conditionals, extracting repeated logic, improving naming, or tightening error handling, and then package those changes into a reviewable pull request.
What this skill actually does
You invoke this when a repository wants steady maintenance improvements without opening a huge rewrite project. It is useful after active feature work, during cleanup windows, or when maintainers want tiny, behavior-preserving refactors to land continuously. It should be used instead of manual refactor brainstorming when the need is routine simplification of fresh code, not architecture redesign. The scope boundary is clear: the workflow targets recent changes, preserves behavior, and validates before opening a PR. It is not a generic product card for GitHub Actions, Copilot, or a programming language.
Integration points include GitHub Actions, the gh CLI, the gh-aw extension, repository labels for automation, and optional CI triggering through GH_AW_CI_TRIGGER_TOKEN. After configuration changes, maintainers run gh aw compile and commit the generated workflow. Because the workflow stays inside a bounded review loop and emphasizes small, test-backed pull requests, it passes the skill-shaped test cleanly.