How does Debugging Master help save LLM tokens?

It uses a three-level system where it first shows a compact timeline (L1), then a structural schema (L2), and only provides full data (L3) upon request or specific JMESPath queries.

Will this tool leave permanent debug logs in my code?

No. All instrumentation is wrapped in specific region markers, allowing the 'cleanup' command to automatically remove all debug-related code once you have found the bug.

What is hypothesis-driven debugging in this context?

It allows you to tag specific instrumentation points with hypothesis IDs (e.g., H1, H2), enabling you to filter log output to see only data relevant to a specific theory about the bug.

Can I use Debugging Master for browser-based applications?

Yes, it includes an HTTP server mode that allows browser environments or isolated containers to send logs via POST requests for centralized analysis.

Debugging Master

Name: Debugging Master
Author: genesiscz

bygenesiscz

•

Analytics & Monitoring

Instruments and analyzes runtime behavior to resolve complex bugs through empirical data observation rather than guesswork.

Debugging Master is a comprehensive runtime observability suite for Claude Code designed to solve bugs that static analysis and code reading cannot explain. It allows developers to instrument their code with targeted logging, performance timers, and state snapshots while maintaining token efficiency through a unique three-level progressive detail system (Compact, Schema, and Full Data). By facilitating hypothesis-driven workflows and session comparisons, it helps isolate environment-specific failures, performance bottlenecks, and data discrepancies across execution layers, all while providing automated cleanup to keep the codebase pristine.

Key Features

01Multi-session comparison to identify execution divergence between passing and failing runs

024 GitHub stars

03High-resolution performance profiling with timer-start/end mechanics

04Hypothesis-driven tagging to filter logs by specific debugging theories

05Targeted runtime instrumentation with region-aware automated cleanup

06Progressive detail levels (L1-L3) to optimize LLM token consumption

Use Cases

01Inspecting large data shapes and API responses without exhausting context windows

02Diagnosing environment-specific failures that occur in CI/CD but pass locally

03Profiling slow operations or identifying hanging execution paths in complex logic

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add genesiscz/genesistools debugging-master

For use in Claude.ai and ChatGPT

Key Features

01Multi-session comparison to identify execution divergence between passing and failing runs

024 GitHub stars

03High-resolution performance profiling with timer-start/end mechanics

04Hypothesis-driven tagging to filter logs by specific debugging theories

05Targeted runtime instrumentation with region-aware automated cleanup

06Progressive detail levels (L1-L3) to optimize LLM token consumption

Use Cases

01Inspecting large data shapes and API responses without exhausting context windows

02Diagnosing environment-specific failures that occur in CI/CD but pass locally

03Profiling slow operations or identifying hanging execution paths in complex logic