How to bring your software delivery workflow into GitHub with agent apps
GitHub introduces Agent Apps to integrate external services like Amplitude, Endor Labs, LaunchDarkly, and PagerDuty directly into its platform, streamlining software delivery workflows without leaving GitHub.
Video
Video available
GitHub has launched Agent Apps to embed tools such as Amplitude, Endor Labs, LaunchDarkly, and PagerDuty directly into its platform, reducing the need to switch between multiple tabs or tools. Developers can now access these services within GitHub’s interface, enabling faster decision-making during pull requests. For example, product insights from Amplitude can be queried directly to assess feature impact before coding begins. This integration aims to centralize workflows and minimize context switching for teams.
The Endor Labs Agent App allows developers to check for vulnerabilities in updated dependencies during pull requests, rather than waiting for CI scans to fail later. By identifying risks proactively, teams can address issues before they escalate. The LaunchDarkly Agent App simplifies feature flag management by automating setup within GitHub, reducing manual coordination via Slack or other tools. These changes are implemented as commits, with human review still required for approvals, maintaining control over rollouts.
The PagerDuty Agent App assesses deployment risks by analyzing active incidents, historical data, and pull request changes to recommend whether to proceed. This routine check replaces ad-hoc reviews only when risks seem high, making risk assessment a standard part of the workflow. The goal is to integrate these tools seamlessly, ensuring developers can focus on coding without juggling external platforms.
Agent Apps are now available in the GitHub Marketplace, allowing organizations to install and enable them for their workflows. GitHub emphasizes that these tools remain unchanged but are now accessible where developers already work. The company provides resources for writing prompts and using Copilot in Java, highlighting the flexibility for enterprise developers to adapt the integration to their needs.