dagger

Why we included this project

Dagger replaces brittle shell scripts and hand-written YAML with a programmable engine for automating software delivery. If you maintain CI pipelines that break whenever someone edits a config file, the appeal is that the same workflow runs identically on a laptop, in a CI server, or in the cloud, with a container runtime as the only hard dependency. Delivery logic is written as typed, composable code through a system API and SDKs in eight languages, and content-addressed caching means changing one file only re-runs the operations that depend on it. Every run emits full OpenTelemetry traces, so debugging a failed pipeline means inspecting a trace instead of wading through log walls. Platform and DevOps engineers who want delivery automation they can reason about will find a lot to like here, and the native LLM support makes it a practical way to run AI agent workflows inside the same pipeline.