Skip to content

🐙 GitHub Integration

Connect AgentOS to your GitHub repositories to track PRs, commits, issues, and contributor velocity in real-time.


🔑 Security & Authentication

AgentOS uses Personal Access Tokens (classic) to communicate directly with GitHub's REST API. Your token is never shared with UnicoLab — it stays encrypted on your local machine.

1

Generate Token

Visit github.com/settings/tokens and create a new Classic token.

2

Select Scopes

Enable the repo scope (covers private repositories, PRs, and issues).

3

Connect

Copy the token and paste it into the Settings > Integrations > GitHub panel in the AgentOS Web UI.


📊 Mapping Your Workspace

Once the connector is active, you can link repositories to your projects.

Multi-Repo Aggregation

Jean-Pierre can track multiple repositories under a single project. This is perfect for microservices or mono-repo workflows where a single "product" spans several codebases.

Project Definition

  • Project: Core Engine
  • Repos: myorg/api, myorg/frontend, myorg/docs
  • Index Depth: 30 Days (Recommended)

🏗️ What Jean-Pierre Indexes

  • Pull Requests: Status, review blockers, and long-running "stale" PRs.
  • Commits: Author velocity, heatmap activity, and codebase churn.
  • Issues: Label-based tracking and milestone progress.
  • Contributors: Active developers and contribution streaks.

⚡ Real-Time Webhooks (Advanced)

By default, AgentOS polls GitHub every few minutes. If you want instant updates, you can expose your local server (e.g., via ngrok) and set up a repository webhook pointing to:

http://your-domain.com:18080/v1/webhook/github


🚀 What GitHub Unlocks

Once connected, these features come alive with your real GitHub data:

Delivery Pulse

Real-time velocity tracking and AI burndown predictions from your commit and PR history. Learn more →

Sprint Forge

AI sprint planning calibrated to your team's real velocity from GitHub activity. Learn more →

Team Pulse

Workload analysis and burnout detection from contributor commit and review patterns. Learn more →

Retro Intelligence

Data-driven retrospectives using PR cycle times, commit patterns, and blocker durations. Learn more →