What AI models does this GitHub workflow automation support?

The skill uses Anthropic's Claude (example shows `claude-3-sonnet-20240229` via `@anthropic-ai/sdk`), though the patterns are designed to work with any AI model API.

How does the automated PR review work?

It triggers on pull request open/synchronize events, diffs the PR against the base branch, sends the diff to an AI model for structured review (including summary, potential issues, suggestions, and security concerns), and posts the review as a PR comment.

What git operations can be automated with this skill?

The skill covers `/rebase` comment-triggered rebasing with force-with-lease, AI-assisted cherry-picking with conflict prediction, and scheduled branch cleanup that identifies branches untouched for 30+ days.

Does this skill include issue triage automation?

Yes—it runs on new issues, sends the title/body to an AI classifier that returns type, severity, area, and suggested labels/assignees, then applies those labels and requests repro steps for under-specified bug reports.

Skill

Automate GitHub Workflows with AI

Name: GitHub Workflow Automation
Availability: OnlineOnly
Author: Antigravity

AI-assisted GitHub workflow automation: PR review, issue triage, smart CI/CD, git operations, and @mention bot commands.

Get skill

Works with githubanthropic

Antigravity

Own this? Claim it

Spark score

out of 100

Updated last month

Version 13.1.0

Models

gemini 2 0

Add to Favorites

Why it matters

Streamline your development lifecycle by automating GitHub workflows. This asset leverages AI to enhance code reviews, triage issues, and integrate AI into your CI/CD pipelines.

Outcomes

What it gets done

Automate Pull Request reviews with AI-generated feedback.

Implement AI-driven issue triage and labeling.

Create and integrate AI into GitHub Actions for CI/CD.

Automate Git operations and smart test selection.

Install

Add it to your toolbox

Run in your project directory:

curl -fsSL https://spark.entire.vc/get/ag-github-workflow-automation | bash

Overview

GitHub Workflow Automation

Patterns for adding AI assistance to GitHub workflows - automated PR review and issue triage, AI risk-gated CI/CD deploys with rollback, git rebase/cherry-pick/branch-cleanup automation, and an @mention bot for on-demand help, all built on GitHub Actions. Use when automating PR reviews, issue triage, AI-gated deployments, or git operations on GitHub; skip it for tasks outside GitHub workflow automation.

What it does

Provides patterns for automating GitHub workflows with AI assistance, inspired by Gemini CLI and modern DevOps practices - covering automated PR reviews, issue triage, AI-validated CI/CD, git operations like rebasing and cherry-picking, and an on-demand @mention bot.

When to use - and when NOT to

Use this skill when automating PR reviews with AI, setting up issue triage automation, creating GitHub Actions workflows, integrating AI into CI/CD pipelines, or automating git operations like rebases and cherry-picks.

Inputs and outputs

Automated PR review workflow triggers on pull_request open/synchronize, diffs the PR against the base branch, sends the diff to an AI model (example uses @anthropic-ai/sdk with claude-3-sonnet-20240229) for a structured review (summary, potential issues, suggestions, security concerns), and posts it as a PR comment via github.rest.pulls.createReview. Reviews can be filtered to specific file extensions and can include full file context, not just the diff.

Issue triage automation runs on issues: opened, sends the title/body to an AI classifier returning JSON (type, severity, area, summary, whether repro steps are present, suggested labels/assignees), applies labels accordingly, and posts a templated request for repro steps on under-specified bug reports. A companion stale-issue workflow (actions/stale@v9) closes inactive issues/PRs after configurable grace periods, exempting labels like pinned, security, and in-progress.

Smart CI/CD patterns include: path-based smart test selection (only run the api, frontend, or database suite based on which paths changed, defaulting to all if nothing specific matches); AI risk assessment before production deploys (returns riskLevel/concerns/recommendations/requiresManualApproval and fails the job via core.setFailed on high risk); and automated rollback via workflow_dispatch that finds the last stable git tag and redeploys it with a Slack notification.

Git operations covered: /rebase comment-triggered PR rebasing with git push --force-with-lease; AI-assisted cherry-picking that predicts conflicts before attempting them and can drive AI-assisted conflict resolution; and scheduled branch cleanup that finds branches untouched for 30+ days and opens a housekeeping issue listing them (with a /keep branch-name escape hatch).

The @ai-helper mention bot responds to PR/issue comments with context-aware answers, supporting commands like @ai-helper explain, review, fix, test, docs, plus /rebase, /update, /approve, /label bug, and /assign @user.

Repository configuration patterns include a CODEOWNERS file scoped by directory/team and branch protection rules configured via the GitHub API (required status checks, required approving reviews, code-owner review requirement, linear history, no force pushes).

Integrations

Built on GitHub Actions (actions/checkout, actions/github-script, actions/stale), the GitHub REST API, an Anthropic SDK client for AI calls, and Slack via slackapi/slack-github-action for notifications. References Google's Gemini CLI GitHub Action as inspiration.

Who it's for

DevOps engineers and maintainers who want to add AI assistance to their GitHub workflow - automated PR review and issue triage, AI-gated deployments, smart test selection, and an @mention bot for on-demand help - on top of standard GitHub Actions infrastructure.

FAQ

Common questions

Discussion