Does this skill work with subagents?

Yes, it includes specific detection for 'Task' tool usage patterns, allowing you to analyze subagent spawning frequency and success rates.

What data does this skill analyze?

It primarily analyzes telemetry data from the 'claude_code_enhanced' job stored in Loki, including tool calls, results, and session metadata.

Can it help with context window issues?

Yes, it monitors context compaction events and high utilization sessions to provide optimization suggestions for managing the context window.

How does it detect workflow inefficiencies?

The skill identifies 'waste' patterns such as multiple reads of the same file, repeated failed operations, and excessive globbing before a read command.

Observability Pattern Detector

Name: Observability Pattern Detector
Author: adaptationio

byadaptationio

分析と監視

Analyzes Claude Code telemetry to identify performance bottlenecks, failure patterns, and workflow inefficiencies.

概要

The Observability Pattern Detector is a specialized tool for Claude Code that automates the recognition of patterns and anomalies within telemetry data. By querying Loki logs, it identifies recurring tool failures, detects slow response sequences, and flags session anomalies like high turn counts or frequent context compactions. This skill is essential for developers looking to optimize their AI workflows, debug complex tool interactions, and reduce operational waste by identifying redundant steps like repeated file reads or inefficient editing patterns. It provides actionable insights through structured reports, helping users refine their prompts and tool usage for maximum efficiency.

主な機能

0 GitHub stars
Long-term trend analysis for tool usage and error rates
Context management monitoring and optimization suggestions
Inefficiency tracking to identify redundant tool calls
Automated failure grouping and impact ranking
Statistical anomaly detection for sessions and response times

ユースケース

Optimizing Claude's token usage by identifying frequent context compactions
Debugging persistent tool errors across different projects
Improving agentic performance by analyzing subagent spawning and success rates

概要

主な機能

0 GitHub stars
Long-term trend analysis for tool usage and error rates
Context management monitoring and optimization suggestions
Inefficiency tracking to identify redundant tool calls
Automated failure grouping and impact ranking
Statistical anomaly detection for sessions and response times

ユースケース

Optimizing Claude's token usage by identifying frequent context compactions
Debugging persistent tool errors across different projects
Improving agentic performance by analyzing subagent spawning and success rates