ColeMurray/background-agents

Background Agents is an open-source background coding agent system that executes development tasks asynchronously while you work on other things.

The tool solves the problem of managing long-running development work by providing a hosted agent that can access full development environments including Node.js, Python, git, browser automation, and VS Code. It works by accepting tasks through multiple interfaces—a web UI, Slack, GitHub PRs, Linear issues, or webhooks—and executing them in sandboxed environments. The agent can spawn parallel sub-tasks in separate sandboxes, create pull requests with proper commit attribution to the user who prompted the work, and run both scheduled automations via cron and event-driven automations triggered by GitHub events, Sentry alerts, or webhooks. It supports multiple AI models including Anthropic Claude, OpenAI, xAI Grok, and OpenCode Zen.

Adoption requires understanding that this tool is designed exclusively for single-tenant deployment where all users are trusted members of the same organization. The system uses a shared GitHub App installation for git operations, meaning authorized users can access any repository the App has access to without per-repository validation. Pull request creation uses the prompting user's GitHub OAuth token when available to ensure proper attribution, but falls back to a shared bot account for users without SCM tokens. This architecture suits internal teams building internal tools where repository access control is already managed at the organizational level, rather than projects requiring multi-tenant isolation or fine-grained per-user repository permissions.

The project shows active development with regular commits addressing core functionality and infrastructure. Work spans across the full stack including the control plane, sandbox environments, web UI, and integrations with external services. The codebase demonstrates attention to security considerations specific to the single-tenant model, with documented token architecture and clear explanations of the authorization flow. Development includes both feature additions and refinement of the deployment and configuration systems.