Telemetry Dashboards for Agent Activity

This build lab focuses on the dashboard that lets humans see what an agent did and why. The goal is not to copy a private machine setup. The goal is to learn the architecture pattern well enough to build a small, classroom-safe version.

An agent dashboard should log events, tool calls, file writes, delegations, todos, failures, and cost counters in a way humans can audit.

Build the small version

1. 1Draw or write a dashboard wireframe with status, recent events, tool calls, file writes, and open todos.
2. 2Mark which parts are user-facing, which parts are internal, and which parts require approval.
3. 3Choose one low-risk workflow and implement only that workflow first.
4. 4Add one failure case before adding a second feature.
5. 5Write a short operator note: what the agent may do, what it must ask about, and what it must never do.

event_log row:
  id
  time
  session_id
  actor
  event_type
  summary
  tool_name
  risk_level
  cost_estimate
  result_status

The big idea: telemetry is not decoration. It is part of the product architecture students need before an agent becomes safe enough to use with real people.