Hybrid Squad

A Hybrid Squad is a team unit that explicitly includes both human engineers and managed AI agents as first-class members. Unlike teams that treat AI as a tool individuals use privately, a Hybrid Squad formalizes agent participation with defined roles, responsibilities, and operational metrics.

The Agentic Development Handbook defines six roles within a Hybrid Squad:

1. Context Architect — writes Live Specs and maintains the Context Index; owns specification quality.
2. Agent Operator — monitors agent execution, responds to Blocker Flags with Rescue Missions, and reviews agent output.
3. Evaluation Engineer — builds and maintains the Eval Harness that validates agent output before human review.
4. Flow Manager — manages pipeline throughput, leads the Daily Flow Sync, and tracks operational metrics on the AgentOps Dashboard.
5. Agent Platform Engineer — provisions and maintains the infrastructure that agents run on, including Ephemeral Workbenches and orchestration tooling.
6. Principal Systems Architect — defines system boundaries, curates Golden Samples, and makes architectural decisions that agents cannot.

The squad scales by expanding agent autonomy boundaries, not headcount:

• 1-2 agents — team members combine roles. A single engineer might serve as both Context Architect and Agent Operator. The Eval Harness is minimal.
• 3-5 agents — dedicated roles emerge. A full-time Agent Operator becomes necessary as concurrent agent sessions increase, and the Evaluation Engineer role separates from general development.
• 5+ agents — all six roles are staffed. The Operator Leverage Ratio becomes a critical metric, and the Flow Manager role is essential for coordinating pipeline throughput.

The key distinction of a Hybrid Squad is that agents are not invisible productivity boosters — they are visible team members whose work is planned, tracked, and reviewed through the same operational processes as human contributions.