Skip to content
SnakeFlow SnakeFlow

GitHub Integration

SnakeFlow integrates deeply with GitHub — Issues, PRs, and CI Actions are managed directly from the sidebar.

Requirements

  • GitHub in your editor — sign in via the Accounts menu (GitHub) for Issues, Pull Requests, and API-based features. The extension uses your editor’s GitHub session.
  • gh CLI — still recommended for some workflows (e.g. team scripts); not required for Issues/PRs or PR AI Feedback (those use the GitHub REST API with your IDE session).
  • For GitHub Actions locally: act + Docker Desktop

GitHub Issues

Features

  • View issues with real-time filters: milestone, assignee, date range, label, full-text search
  • Create new issues (optional AI-assisted title via Google Gemini)
  • Edit title, body, assignees, milestone, labels
  • Close issues
  • Add comments
  • Create a branch from an issue — names the branch issue-{number}-{title-slug}
  • Send to Cursor Chat — sends the full issue context for AI assistance

Command: Ctrl+Alt+G — Create Issue | Sidebar → GitHub Issues panel

Filters

The sidebar shows up to 50 issues. Filter by:

  • Milestone — select from your repo’s milestones
  • Assignee — filter to your issues or a specific user
  • Date — created after / before a date
  • Search — full-text search in title and body

AI Title Generation (Gemini)

"devManager.gemini.apiKey": "AIza...",
"devManager.gemini.model": "gemini-2.5-flash"

When configured, SnakeFlow calls Gemini to suggest a concise issue title based on your description.

Get a free API key at aistudio.google.com.

Voice Dictation

Click the 🎙 button next to Description (or Add comment in Edit Issue) to dictate text using your microphone. Gemini AI transcribes the audio and inserts the result directly into the field — no copy-paste required.

See Voice Dictation → for setup and language settings.

Send to Chat

"devManager.issueToChat.prompt": "Analyze this issue and suggest an implementation plan."

Pull Requests

View open PRs in the sidebar with review status, CI status, and author.

Features

  • PRs grouped by status: needs review, approved, changes requested
  • Open PR on GitHub with one click
  • Send PR to Cursor Chat — sends the diff + description for AI code review

Custom Chat Prompt

"devManager.prToChat.prompt": "Review this PR focusing on security, performance, and test coverage."

PR AI Feedback

Build a Markdown report for one open pull request — PR metadata, CI check results, bot reviews and inline comments (e.g. Copilot, CodeRabbit, GitHub Actions bots), plus an excerpt from failed job logs when available. Intended as a single file you can paste into an AI chat or archive for review.

How to run

  1. Select your SnakeFlow project and ensure GitHub is signed in in the editor.
  2. Command Palette → SnakeFlow: PR AI Feedback (or open it from the main menu).
  3. Pick an open PR from the list (your current branch is marked when it matches a PR head).
  4. Wait for the notification — the report opens in the editor.

Where the file is saved

The report is written under your project root, inside the editor configuration folder (the same parent folder that holds skills or workspace metadata — often .cursor in Cursor or .vscode in VS Code / Antigravity), in a subfolder named pr-feedback/. Filename pattern: pr-{number}-{YYYY-MM-DD-HH-mm}.md. Add pr-feedback/ to .gitignore if you do not want reports committed.

Requirements: GitHub origin remote must point at github.com so the repository can be detected. Private repos need token scopes that allow reading PRs, checks, and comments (the default repo flow from the editor is usually enough).

GitHub Actions (CI)

Run GitHub Actions workflows locally using act — without pushing to GitHub.

Install act

Terminal window
# Windows
winget install nektos.act


# macOS
brew install act


# Linux
curl -s https://raw.githubusercontent.com/nektos/act/master/install.sh | sudo bash

Requires Docker Desktop to be running.

Usage

Main menu → CI / Run GitHub Actions Locally

"devManager.ci.workflowsPath": ".github/workflows"