Worker Agent

• Key points are in the visual diagram above.

Specialized workers with focused context windows outperform generalist agents on domain tasks and are cheaper to eval and iterate independently.

Workers are specialty contractors on a build site — the electrician doesn't pour concrete but delivers a certified wiring package to the general contractor.

Scenario

SQL Worker Agent: receives natural-language question + schema, returns query results or error — never decides what question to ask next.

Solution

In Multi-Agent Systems, apply Worker Agent to this scenario: SQL Worker Agent: receives natural-language question + schema, returns query results or error — never decides what question to ask next. Identify the inputs, run the technique, validate the output, and note one thing you would monitor in production.

Do this before moving to the next module — reading alone is not enough.

Open the Code Walkthrough below and run it locally. Change one parameter related to Worker Agent (e.g. model, temperature, top_k, or tool name), observe the difference in output, and write 2–3 sentences explaining what changed.

Highlighted lines show where Worker Agent happens in the code.

1sql_worker = Agent(  # key line for Worker Agent2    instructions="Execute SQL only. Return JSON {rows, error}.",3    tools=[run_sql],4    output_type=SQLResult,5)

• Treating Worker Agent as a black box without evaluation
• Ignoring cost and latency in production
• Skipping error handling for worker agent

Quick recap — the most important points from this module.