gpd-debug

Diagnoses physics calculation errors and generates automated root-cause reports.

1|Updated Mar 29, 2026

Search Tags:#parallel processing #error diagnosis #systematic investigation #computational physics #physics debugging

Ready to run

npx skills add https://github.com/CharGrnmn/roomtemp-superconductor-gpd --skill gpd-debug-chargrnmn

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: gpd-debug
Source: https://github.com/CharGrnmn/roomtemp-superconductor-gpd/tree/main/.agents/skills/gpd-debug
Command: npx skills add https://github.com/CharGrnmn/roomtemp-superconductor-gpd --skill gpd-debug-chargrnmn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires gpd, gpd-debugger, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the complex and nuanced debugging of physics calculations, providing a systematic approach with persistent state to investigate and resolve issues in calculations.

Core Features & Use Cases

Systematic Debugging: Utilizes a structured approach to diagnose research problems and find root causes in physics calculations.
Persistent State: Maintains state across context resets, ensuring that investigations can be carried out in parallel with minimal overhead.
Subagent Investigation: Spawns investigation agents for each issue, allowing for independent and parallel investigation.
Mode Detection: Detects invocation context (batch or interactive) to determine the mode of operation.
Quick Triage: Offers a set of common patterns and causes for discrepancies, allowing for immediate resolution of many issues.
Process Automation: Automates the process of parsing gaps, reporting diagnosis plans, spawning agents, collecting results, updating verification, and reporting diagnosis results.

Quick Start

Use the gpd-debug skill to investigate an issue in your physics calculation by providing a brief description of the problem.

Frequently Asked Questions about gpd-debug

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

FAQPage Schema

How do I debug numerical instabilities and convergence failures in physics calculations?▼

Debugging physics calculation errors systematically involves diagnosing numerical instabilities and convergence failures through persistent state tracking. This approach spawns parallel investigation agents to identify root causes like wrong approximations and unphysical results.

What is the best way to investigate unphysical results and sign errors in computational physics?▼

Investigating unphysical results and sign errors in computational physics requires a systematic debugging approach. It maintains persistent state across context resets to trace theoretical inconsistencies and automate the collection and verification of diagnosis results.

Do I need the GPD runtime and gpd-debugger agent to diagnose physics calculation errors?▼

Yes, diagnosing physics calculation errors with this systematic approach requires both the GPD runtime and the gpd-debugger agent. These dependencies enable Python orchestration and the spawning of independent investigation agents for parallel processing.

Can I run parallel investigations for physics debugging across context resets?▼

Yes, you can run parallel investigations for physics debugging across context resets. The system maintains persistent state and spawns independent subagents for each issue, ensuring investigations proceed with minimal overhead when the invocation context resets.

How do I quickly triage common patterns causing discrepancies in physics calculations?▼

Quick triage of common patterns causing discrepancies in physics calculations is achieved by matching issues against a set of known causes. This allows for immediate resolution of many numerical instabilities and theoretical inconsistencies without spawning deep investigation agents.

When should I avoid automated subagent investigation for physics calculation errors?▼

You should avoid automated subagent investigation for physics calculation errors when the issue matches common triage patterns for immediate resolution. For complex numerical instabilities or theoretical inconsistencies, the automated parallel investigation provides better root cause diagnosis.