Concepts › DAG Pipelines

DAG Pipelines

ADA pipelines are directed acyclic graphs. Each node is a phase, each edge a dependency. Phases without dependencies execute in parallel.

Topologies

sequential

Phases A → B → C. Each phase depends on the previous one. Use case: DB migrations then backend then frontend.

parallel_mesh

Phases A, B, C are independent. All launched simultaneously. Use case: tasks with no cross-dependencies.

hybrid

Mixed sequential/parallel. Phases A+B run in parallel, then C depends on both A and B. Most common case.

Pipeline structure

{
  "pipeline": "feature-fullstack",
  "topology": "hybrid",
  "phases": [
    { "id": "db-schema",     "agent": "db",     "tier": 1, "deps": [] },
    { "id": "tdd-specs",     "agent": "tester", "tier": 1, "deps": ["db-schema"] },
    { "id": "backend-impl",  "agent": "nestjs", "tier": 3, "deps": ["db-schema", "tdd-specs"] },
    { "id": "frontend",      "agent": "nextjs", "tier": 2, "deps": ["backend-impl"] }
  ]
}