Chapter 7: Telemetry, Cost, and Team Governance

Welcome to Chapter 7: Telemetry, Cost, and Team Governance. In this part of HumanLayer Tutorial: Context Engineering and Human-Governed Coding Agents, you will build an intuitive mental model first, then move into concrete implementation details and practical production tradeoffs.

Team-scale coding-agent programs need quality, cost, and governance telemetry to remain effective.

Key Metrics

Area	Metrics
quality	accepted patch rate, regression rate
efficiency	time-to-first-useful-diff
cost	token spend per completed task
governance	policy violations, manual overrides

Summary

You now have a metric framework for sustainable team operations.

Next: Chapter 8: Production Rollout and Adoption

Source Code Walkthrough

claudecode-go/types.go

The claudecode-go/types.go file defines the structured event types emitted by the Claude Code subprocess — tool calls, results, usage metrics, and agent outputs. These typed events are the raw telemetry stream from which the quality and cost metrics described in this chapter (accepted patch rate, token spend per task) are derived.

humanlayer.md

The humanlayer.md SDK reference documents the governance hooks — policy violation callbacks, manual override logging, and audit trail capture — that feed the governance metrics (policy violations, manual overrides) in the metrics table of this chapter.

How These Components Connect

flowchart TD
    A[claudecode-go/types.go] -->|typed events| B[Telemetry stream]
    B -->|token counts| C[Cost metrics: spend per task]
    B -->|tool call/result pairs| D[Quality metrics: patch rate]
    B -->|time deltas| E[Efficiency metrics: time-to-diff]
    F[humanlayer.md governance hooks] -->|override events| G[Governance metrics]
    C --> H[Team dashboard]
    D --> H
    E --> H
    G --> H