Pipeline Chains

Sequential stages where each phase’s output feeds the next, with context reset between stages. How it works:

Create tasks with dependencies: each stage is blocked by the previous
Each handoff resets context — research noise doesn’t leak into implementation
Each stage writes structured output that the next stage consumes

Best for: research → implementation flows, multi-stage migrations, any workflow where earlier context would bias later steps. A typical four-stage pipeline:

Research (Haiku) → Plan (Opus) → Implement (Sonnet) → Validate (Sonnet, read-only)

Each stage uses the cheapest model that can do the job — see cost and model routing. Division of labor:

Stage	Orchestrator	Agents
Research	Coordinates topics	3-4 parallel researchers
Planning	Manages revision loop	Planner + checker until pass
Execution	Groups into waves, tracks	Executors in parallel, fresh context each
Verification	Routes next step	Verifier checks code against goals

The orchestrator never does heavy lifting — it spawns, waits, integrates, routes.

Examples in the wild

Example	What it shows
cyber-defense-team	4-stage pipeline: log-ingestor → anomaly-detector → risk-classifier → threat-reporter. Each agent reads the previous stage’s JSON and writes new JSON. Includes cost estimates per stage.
gsd:quick	Composable pipeline: `--discuss`, `--research`, `--full` each add a stage. Omit what you don’t need.

Daily Workflow

Autonomous Work

Multi-Agent

Operations

Examples in the wild

​Examples in the wild

Examples in the wild