skills.rest
ZESCODE
ZESCODECOMMUNITY
Explore all skills by this author →
GitHub

ZES-systematic-debugging

Orchestrate a multi-agent debugging workflow for root cause analysis and fixes.

Updated Jul 8, 2026
Search Tags:#automation#debugging#multi-agent#cdp#root-cause-analysis#browser-harness
One-click install
npx skills add https://github.com/ZESCODE/Zes-Orchestration-System --skill zes-systematic-debugging
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ZES-systematic-debugging
Source: https://github.com/ZESCODE/Zes-Orchestration-System/tree/main/.agents/skills/ZES-systematic-debugging
Command: npx skills add https://github.com/ZESCODE/Zes-Orchestration-System --skill zes-systematic-debugging

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill addresses the inefficiency of manual debugging by coordinating three specialized AI agents to simultaneously gather evidence, analyze root causes, and implement verified fixes.

Core Features & Use Cases

  • Parallel Debugging Pipeline: Utilizes an Explorer for evidence, an Analyzer for root cause, and a Fixer for implementation.
  • Advanced Browser Integration: Supports CDP direct access, interactive browser-harness sessions, and autonomous browser-use exploration.
  • Use Case: When a complex dashboard error occurs, the system automatically captures console logs, identifies the architectural failure, and generates a precise code patch with rollback commands.

Quick Start

Run the zes debug command followed by your error message and the target directory to initiate the automated 3-agent investigation.

Frequently Asked Questions about ZES-systematic-debugging

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

FAQPage Schema
How do I automate root cause analysis for complex web dashboard errors?▼

Automated root cause analysis for web dashboard errors is handled by a parallel 3-agent pipeline that gathers evidence, analyzes architectural failures, and generates verified code patches with rollback commands.

Can I use CDP to capture console logs during a multi-agent debugging session?▼

CDP direct access is supported for capturing console logs during automated debugging sessions. The system uses an Explorer agent to gather browser evidence and an Analyzer agent to determine the root cause.

What is the best way to coordinate parallel AI agents for software defect resolution?▼

Coordinating parallel AI agents for software defect resolution involves assigning specialized roles: an Explorer for evidence collection, an Analyzer for root cause identification, and a Fixer for implementing verified code patches.

Do I need specific LLM providers to run the systematic diagnostic pipeline?▼

The systematic diagnostic pipeline requires integration with Groq, OpenRouter, and LLM7 providers to execute the multi-agent workflow and perform automated bug investigation across local codebases.

How to start an automated debugging investigation from a specific directory?▼

To start an automated debugging investigation, run the zes debug command followed by your error message and the target directory to initiate the 3-agent evidence gathering and fix implementation pipeline.

Does browser-harness support autonomous exploration for local codebase debugging?▼

Browser-harness supports interactive sessions and autonomous browser-use exploration to gather evidence across web dashboards and local codebases, feeding the root cause analysis pipeline for bug resolution.

Related Skills

Explore More Software Engineering Skills→
jamesrochabrunjamesrochabrun
194

llm-router

Route LLM requests to OpenAI, Grok, Groq, DeepSeek, and OpenRouter via SwiftOpenAI-CLI.

Community
Intermediate
Ard2pArd2p

aif-evolve

Analyze patches and code patterns to update aif-* skill-context policies.

Community
Advanced
twjarviszyra-webtwjarviszyra-web
3

dogfood

Test web applications and capture reproducible bug evidence with screenshots.

Community
Advanced
cspergelcspergel

Codebase Sweep

Orchestrate autonomous sweeps to identify cross-cutting issues across ForgePlan codebases.

Community
Advanced