debug

Identify and remediate software errors using a structured debugging workflow with verified tool outputs.

2|1|Updated Mar 8, 2026

Search Tags:#troubleshooting #verification #debug #engineering #toolchain #root-cause #incident-response

Ready to run

npx skills add https://github.com/duggal1/Sapphire-cli --skill debug-duggal1

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: debug
Source: https://github.com/duggal1/Sapphire-cli/tree/main/internal/skills/bundled/debug
Command: npx skills add https://github.com/duggal1/Sapphire-cli --skill debug-duggal1

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill provides a disciplined, tool-backed approach to identifying, isolating, and remediating software errors by requiring traceable evidence from executed steps and results.

Core Features & Use Cases

Structured debugging workflow: reproduce failure boundaries, collect tool outputs, and validate fixes with auditable traces.
Comprehensive tool-chain execution: integrate web searches, codebase reading, and terminal commands to surface root causes.
Use Case: When a flaky deployment occurs, run the full tool-chain to verify the root cause and confirm a reproducible fix with logs.

Quick Start

Begin a guided debugging session by outlining the failure boundary and requesting a verified plan with step-by-step tool outputs.

Frequently Asked Questions about debug

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

FAQPage Schema

How do I debug software errors using verifiable evidence instead of guessing?▼

To debug software errors with verifiable evidence, apply a structured workflow that reproduces failure boundaries, collects terminal and codebase outputs, and validates fixes using auditable traces. This enforces a full tool-chain mandate for root-cause isolation.

What is the best way to troubleshoot a flaky deployment and confirm the root cause?▼

Troubleshooting a flaky deployment requires running the full tool-chain—web search, codebase reading, terminal execution, and file inspection—to verify the root cause. You confirm a reproducible fix by validating the remediation with collected logs.

How do I isolate failure boundaries in a complex codebase during incident response?▼

Isolating failure boundaries during incident response involves executing terminal commands and reading the codebase to trace errors. You reproduce the issue, collect tool outputs, and validate the fix with auditable evidence.

Can I use terminal execution and file inspection to verify a software fix?▼

Yes, you can use terminal execution and file inspection to verify a software fix. The debugging workflow mandates tool execution to drive evidence-based conclusions, ensuring fixes are validated with auditable outputs.

Does structured troubleshooting work for reproducing runtime environment issues?▼

Structured troubleshooting works across codebases and runtime environments by reproducing issues and collecting tool outputs. It enforces a full tool-chain mandate, including web fetch and codebase reading, to surface root causes.

What are the limitations of debugging without a full tool-chain mandate?▼

Debugging without a full tool-chain mandate limits your ability to isolate failure boundaries and validate fixes with auditable traces. It skips necessary evidence collection from web searches, terminal execution, and file inspection.