telemetry-insights

Analyze Blueplane telemetry from Claude Code and Cursor sessions for usage patterns and token economics.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/blueplane-ai/bp-telemetry-core --skill telemetry-insights
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: telemetry-insights
Source: https://github.com/blueplane-ai/bp-telemetry-core/tree/main/.claude/skills/telemetry-insights
Command: npx skills add https://github.com/blueplane-ai/bp-telemetry-core --skill telemetry-insights

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns raw AI-assistant telemetry into actionable metrics and recommendations so developers can measure efficiency, detect wasted effort, and improve prompting and workflow patterns without manual data wrangling.

Core Features & Use Cases

  • Raw metrics extraction: aggregates flow, prompting, token counts, patches, and behavioral signals for session or time-window analyses.
  • Derived insights & scores: computes efficiency, context sufficiency, persistence, and workflow quality indicators and predicts stuckness or task difficulty.
  • Platform-aware analysis: supports Claude Code trace inspection (including compressed event blobs and tool-usage extraction) and Cursor session heuristics with workspace mapping.
  • Use case: quantify a week's AI-assisted productivity, identify trial-and-error churn on specific files, and get prioritized recommendations to reduce token cost and improve prompt quality.

Quick Start

Analyze my last coding session and summarize token efficiency, key workflow patterns, and three concrete recommendations.

Frequently Asked Questions about telemetry-insights

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze AI-assisted coding telemetry to measure token efficiency and workflow churn?

AI-assisted coding telemetry analysis extracts usage patterns, token counts, and workflow churn from Claude Code and Cursor sessions. It computes derived efficiency scores and detects prompting behavior to pinpoint wasted effort and reduce token cost.

Can I extract metrics from both Claude Code and Cursor session histories?

Yes, telemetry extraction supports both Claude Code trace inspection, including compressed event blobs and tool-usage detection, and Cursor session heuristics with workspace mapping to aggregate cross-platform behavioral signals.

What is the best way to quantify a week of AI-assisted coding productivity?

Quantifying AI-assisted coding productivity involves running time-windowed telemetry analysis on session databases. It aggregates flow, prompting, and patch behavior to generate derived insights and prioritized workflow recommendations.

Do I need a local SQLite database to analyze my coding session traces?

Yes, you need local access to the Blueplane telemetry storage, specifically the SQLite database at ~/.blueplane/telemetry.db. The analysis decodes compressed event blobs and joins session and trace tables directly from this local source.

How does telemetry analysis detect trial-and-error churn on specific files?

Telemetry analysis detects trial-and-error churn by classifying patch behavior and cross-referencing git metadata. It identifies repetitive prompting patterns and applies heuristics to predict task difficulty and context sufficiency.