How does Enhanced Telemetry differ from default Claude Code telemetry?

While default telemetry provides basic OTEL metrics, Enhanced Telemetry utilizes every available hook to capture granular data like tool-to-tool sequences, subagent types, and specific context compaction triggers.

Can I track token usage per repository?

Yes, the skill includes repository analytics that capture tool usage, prompt volume, and estimated token consumption categorized by the specific project directory.

What infrastructure is required to view the telemetry data?

The skill is designed to work with a Loki instance for log aggregation and Grafana for visualization, though the Python scripts can be adapted for other logging backends.

Does this skill require Python to run?

Yes, the telemetry hooks are implemented via a Python 3 script that must be accessible in your system PATH.

Enhanced Telemetry

Name: Enhanced Telemetry
Author: adaptationio

byadaptationio

分析与监控

Captures comprehensive observability metrics across the entire Claude Code lifecycle using ten specialized hooks for deep session and tool analysis.

关于

Enhanced Telemetry is a robust observability skill designed for developers who need deeper insights into how Claude Code interacts with their codebase. By instrumenting all ten available lifecycle hooks, this skill tracks everything from session duration and prompt patterns to specific tool sequences and subagent performance. It provides critical data on context window utilization and token accumulation, allowing users to optimize their AI workflows, debug tool failures more effectively, and monitor repository-level activity via integrated Grafana dashboards and Loki logging.

主要功能

Complete instrumentation of all 10 Claude Code hooks for maximum visibility
Real-time context window tracking with token estimation and compaction event logs
Subagent monitoring to track the lifecycle and success rates of spawned agents
Detailed tool analytics including call frequency, error rates, and sequence mapping
0 GitHub stars
Pre-configured LogQL queries and Grafana dashboard support for immediate visualization

使用场景

Debugging complex tool sequences and identifying recurring errors in automated workflows
Analyzing team-wide prompt patterns to refine engineering best practices
Auditing AI performance and token consumption across large-scale development projects

关于

主要功能

Complete instrumentation of all 10 Claude Code hooks for maximum visibility
Real-time context window tracking with token estimation and compaction event logs
Subagent monitoring to track the lifecycle and success rates of spawned agents
Detailed tool analytics including call frequency, error rates, and sequence mapping
0 GitHub stars
Pre-configured LogQL queries and Grafana dashboard support for immediate visualization

使用场景

Debugging complex tool sequences and identifying recurring errors in automated workflows
Analyzing team-wide prompt patterns to refine engineering best practices
Auditing AI performance and token consumption across large-scale development projects